- Email: [email protected]
Injured primary sensory neurons switch phenotype for brain-derived neurotrophic factor in the rat
Pergamon PII:
Neuroscience Vol. 92, No. 3, pp. 841–853, 1999 Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00 S0306-4522(99)00027-5
INJURED PRIMARY SENSORY NEURONS SWITCH PHENOTYPE FOR BRAIN-DERIVED NEUROTROPHIC FACTOR IN THE RAT X.-F. ZHOU,*† E. T. CHIE,* Y.-S. DENG,* J.-H. ZHONG,* QING XUE,* R. A. RUSH* and C. J. XIAN‡ *Department of Human Physiology and Centre for Neuroscience, Flinders University of South Australia, GPO Box 2100, Adelaide 5001, Australia ‡Child Health Research Institute, North Adelaide 5006, Australia
Abstract—Peripheral nerve injury results in plastic changes in the dorsal root ganglia and spinal cord, and is often complicated with neuropathic pain. The mechanisms underlying these changes are not known. We have now investigated the expression of brain-derived neurotrophic factor in the dorsal root ganglia with histochemical and biochemical methods following sciatic nerve lesion in the rat. The percentage of neurons immunoreactive for brain-derived neurotrophic factor in the ipsilateral dorsal root ganglia was significantly increased as early as 24 h after the nerve lesion and the increase lasted for at least two weeks. The level of brain-derived neurotrophic factor messenger RNA was also significantly increased in the ipsibut not contralateral dorsal root ganglia. Both neurons and satellite cells in the lesioned dorsal root ganglia synthesized brain-derived neurotrophic factor messenger RNA after the nerve lesion. There was a dramatic shift in size distribution of positvie neurons towards large sizes seven days after sciatic nerve lesion. Morphometric analysis and retrograde tracing studies showed that no injured neurons smaller than 600 mm 2 were immunoreactive for brain-derived neurotrophic factor, whereas the majority of large injured neurons were immunoreactive in the ipsilateral dorsal root ganglia seven days postlesion. The brainderived neurotrophic factor-immunoreactive nerve terminals in the ipsilateral spinal cord were reduced in the central region of lamina II, but increased in more medial regions or deeper into laminae III/IV. These studies indicate that sciatic nerve injury results in a differential regulation of brain-derived neurotrophic factor in different subpopulations of sensory neurons in the dorsal root ganglia. Small neurons switched off their normal synthesis of brain-derived neurotrophic factor, whereas larger ones switched to a brain-derived neurotrophic factor phenotype. The phenotypic switch may have functional implications in neuronal plasticity and generation of neuropathic pain after nerve injury. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: BDNF, in situ hybridization, spinal cord, RT–PCR, sensory neurons.
Sciatic nerve injury results in distinct changes in sensory neurons of the dorsal root ganglia (DRGs) and their terminals in the spinal cord. In the adult, sciatic nerve lesion leads to a reduction in substance P (SP) and calcitonin gene-related peptide (CGRP), neurotransmitters for small nociceptive sensory neurons, and an increase in neuropeptide Y and vasoactive intestinal polypeptide in large sensory neurons. 17 Transection of the sciatic nerve triggers sprouting of myelinated sensory afferents from deep laminae into lamina II of the spinal cord, where the nociceptive fibres with high thresholds normally project. 49–51 A conditioning lesion of the sciatic nerve confers on sensory neurons an enhanced regenerative response to a secondary injury. 27,36 In †To whom correspondence should be addressed. Abbreviations: BDNF, brain-derived neurotrophic factor; CGRP, calcitonin gene-related peptide; CTB, cholera toxin B; DRG, dorsal root ganglion; GAPDH, glyceraldehyde-3phosphate dehydrogenase; IR, immunoreactive; NGF, nerve growth factor; NT-3, neurotrophin-3; PCR, polymerase chain reaction; RT, reverse transcription; SP, substance P; SSC, standard saline citrate. 841
addition to these morphological changes, nerve injury can induce clinical syndromes characterized by hyperalgesia and allodynia. 18 The mechanisms underlying these changes after nerve injury, however, are not known. Sensory neurons with unmyelinated or small-diameter myelinated fibres express TrkA and are dependent on nerve growth factor (NGF) for their survival. 23 Up to 70% of these neurons projecting to laminae I and II of the spinal cord are lost following gene deletion of TrkA or NGF, or upon administration of NGF antibody to developing animals. 38,40 These animals display hypoalgesia. 24 In contrast, large neurons expressing TrkC and containing neurotrophin-3 (NT-3) 7,30 are supported by NT-3 for their survival during development. 22,53 NT-3 null mutant mice lack proprioceptive and mechanoceptive functions, and the number of Merkel cells in touch domes is reduced. 2 Brain-derived neurotrophic factor (BDNF) and TrkB are normally synthesized in 20– 30% of all spinal sensory neurons and confined to those with small to medium diameters. 14,30 Deletion of the BDNF gene results in the reduction of sensory
842
X.-F. Zhou et al.
neurons in both spinal and cranial ganglia. The percentage of neuron loss varies between 30% in DRGs and 80% in vestibular ganglia. 12 BDNF is required for the normal function of slowly adapting mechanoceptive neurons, but not for their survival. 6 Although BDNF in sensory neurons is anterogradely transported to their peripheral and central targets, 57 it may also function as an autocrine neurotrophic factor. 1 A subpopulation of small nociceptive neurons which do not express any Trk receptors but express c-Ret and bind the lectin, IB4, depends on glial cell line-derived neurotrophic factor for survival. 29,41 It is likely that neurotrophic factors play important roles in neuronal plasticity and neuropathic pain after injury. Recently, we have found that satellite cell-derived NGF and NT-3 are involved in sympathetic sprouting in DRGs after nerve damage. 53 The low-affinity NGF receptor (p75) is also implicated in the formation of baskets of sympathetic nerves in DRGs. 35,55 NGF, but not BDNF or NT-3, delivered intrathecally, blocks the sprouting of myelinated afferents in the spinal cord after sciatic nerve lesion. 5 In a previous study, we found that about 25% of small to medium-sized sensory neurons in the normal rat lumbar DRGs are immunoreactive for BDNF, 57 which is consistent with the percentage of neurons containing mRNAs for BDNF. 14 Recently, it has been shown that peripheral nerve lesion results in an enhanced transport of both exogenous and endogenous BDNF in sensory neurons. 47 In the present study, we sought to determine what changes in BDNF expression occur in primary sensory neurons following sciatic nerve injury in rats, using a variety of techniques.
EXPERIMENTAL PROCEDURES
Animals Adult Sprague–Dawley rats, male and female (bred in Flinders University), were used under the guidelines of the National Health and Medical Research Council of Australia, and approved by the Animal Welfare Committee of Flinders University of South Australia. All efforts were made to minimize animal suffering and to reduce the number of animals.
Surgery Rats were anaesthetized with inhalational halothane, and the left sciatic nerves in the mid-thigh region were ligated and distally transected. Animals were allowed to survive for up to 14 days after lesion (n 5 for immunohistochemistry and n 3 for in situ hybridization for each time-point). Some rats without surgery were used as controls. Fresh L4 and L5 DRGs (n 3 or 4 for each time-point) were dissected bilaterally, frozen in liquid nitrogen and stored at 2708C prior to BDNF mRNA quantification. The remaining animals were perfused with Zamboni’s fixative and prepared for BDNF immunohistochemistry or BDNF in situ hybridization.
Immunohistochemistry of brain-derived neurotrophic factor BDNF immunohistochemistry was performed as described previously. The polyclonal rabbit antibody to BDNF made against recombinant human BDNF has been characterized previously and recognizes BDNF only, without cross-reactivity to other neurotrophins. 57 Rats were perfused with modified Zamboni’s fixative containing 2% formaldehyde and 15% saturated picric acid after an overdose of pentobarbital. After dissection, DRGs and the spinal cord were sectioned on a Cryostat microtome at 30 mm. Free-floating sections were treated with 0.3% hydrogen peroxide in 50% ethanol, blocked in 20% normal horse serum and incubated in rabbit polyclonal antibodies against recombinant human BDNF at a concentration of 1 mg/ml overnight. After extensive washing in phosphate-buffered saline (pH 7.4) containing 0.1% Tween 20, the sections were incubated in biotinylated secondary antibodies to rabbit immunoglobulin G, followed by incubation with the avidin–biotin–peroxidase complex reagent (ABC Kit, Vector Labs). Sections were developed in 0.05% diaminobenzidine containing 0.01% hydrogen peroxide and observed under a light microscope. All positive and negative neurons were counted from three randomly selected sections of L5 DRG using a Normaski optic setting. Only neurons with a visible nucleus were counted. The percentage of BDNF-immunoreactive (IR) neurons was then calculated for each animal, and the means from each group of animals were statistically compared using Student’s t-test.
Quantification of messenger RNA
brain-derived
neurotrophic
factor
The procedures for total RNA extraction and reverse transcription–polymerase chain reaction (RT–PCR) have been described previously. 54 In brief, total RNA was prepared from fresh rat ganglia (n 3) according to the acid guanidinium thiocyanate–phenol–chloroform method 9 using an RNA isolation kit (Advanced Biotechnologies, Leatherhead, U.K.). The RNA was treated with DNAase (Promega, U.S.A.) to remove possible contaminating genomic DNA and then subjected directly to first-strand cDNA synthesis by incubation with oligo dT15 and AMV reverse transcriptase. PCR primers for BDNF and GAPDH cDNAs were designed corresponding to the coding region of the genes as follows: BDNF primers, sense 5 0 TCCCTGGCTGACACTTTTGAG-3 0 and antisense 5 0 CTATCCTTATGAACCGCCAGC-3 0 ; GAPDH primers, sense 5 0 -TGCTGGTGCTGAGTATGTCG-3 0 and antisense 5 0 -GCATGTCAGATCCACAACGG-3 0 . PCR was performed in a 30-ml volume containing thermostable DNA polymerase (Advanced Biotechnologies, U.K.) on a Perkin DNA Thermal Cycler (Perkin Elmer, U.S.A.). All samples were heated at 958C for 2 min and amplified in cycles at 958C for 30 s, 588C for 30 s and 728C for 30 s. The last cycle was followed by a final incubation at 728C for 10 min. The house-keeping gene GAPDH was amplified in parallel to serve as an internal control. PCR cycle numbers for BDNF and GAPDH were optimized into the linear range. The appropriate concentration of cDNA was determined by a preliminary PCR and adjusted for quantitative PCR. The PCR products were electrophoresed and stained with ethidium bromide on a 1.5% agarose gel. The gel was captured as a digital image and analysed using a FluorImager 595 and quantitation software (Molecular Dynamics, U.S.A.). This analysis provided the same results as achieved with a previous method using radioactive hybridization. 56 The ratio of BDNF to GAPDH mRNAs was calculated from the fluorescent values of the PCR products. The mRNA level for each ganglion was expressed as a percentage of the mRNA level in the normal control ganglia.
Injured sensory neurons switch phenotype for BDNF
In situ hybridization In situ hybridization was used to localize BDNF mRNA expression in the tissues using digoxigenin-labelled riboprobes. BDNF antisense and sense riboprobes were made from a ribovector pGEM-3Zf(1) containing 800-bp rat BDNF after being linearized by EcoR I and Hind III and using SP6 and T7 RNA polymerase, respectively, using a digoxigenin-RNA labelling kit (Boehringer Mannheim, Mannheim, Germany). Paraffin sections of 4 mm were cut from tissue specimens, mounted on 3-aminopropyltriethoxysilane-coated glass slides, dewaxed, hydrated and treated with 0.2 N HCl for 20 min at room temperature. Sections were then permeabilized with 1 mg/ml proteinase K (Sigma) for 20 min at 378C. After postfixation for 5 min with 2% paraformaldehyde, sections were neutralized for 10 min with 0.2% glycine in phosphate-buffered saline. Following washing and dehydration, sections were covered with 25 ml hybridization mix (pre-heated for 5 min at 858C), containing 0.5 mg/ml sense (used as a negative control) or antisense probe, 50% formamide, 10% dextran sulphate, 0.05% Triton X-100, 500 mg/ml herring sperm DNA (Boehringer Mannheim), 0.05% polyvinylpyrrolidone and 5 × SSC (750 mM NaCl and 75 mM sodium citrate, pH 7.0). Sections were covered with glass coverslips and incubated in a humidified chamber for 18 h at 518C. Post-hybridization washes were undertaken at room temperature as follows with gentle shaking: once in 2 × SSC solution for 5 min to remove coverslips; twice in 2 × SSC for 15 min; and once in 1 × SSC for 10 min. After blocking of non-specific binding sites with 1% blocking reagent (Boehringer Mannheim) for 30 min, the sections were incubated for 30 min at 378C with an alkaline phosphatase-coupled sheep anti-digoxigenin immunoglobulin G (1:450; Boehringer Mannheim) and a colour reaction developed by incubating in nitroblue tetrazolium/X-phosphate substrate for 18 h in the dark according to the supplier’s instructions (Boehringer Mannheim). Sections were washed with water, counterstained lightly with 0.05% Methyl Green for 5 min and mounted with glycerol gel (Dako). Combination of retrograde labelling and brain-derived neurotrophic factor immunohistochemistry FluoroGold labelling. The left sciatic nerve in three rats was transected without ligation and the proximal stumps were soaked in 4% FluoroGold solution for 5 min. All lesioned neurons in L4 and L5 DRGs were labelled using this method. In control rats (n 2), FluoroGold (1 ml) was injected into the left sciatic nerve without lesion. The rats were allowed to survive for one week before BDNF immunohistochemistry on the DRGs. Cholera toxin B–rhodamine labelling. To label myelinated afferent neurons, cholera toxin B (CTB)–rhodamine tracer was used. After sciatic nerve lesion, the rats were reanaesthetized with inhalational halothane and the lesioned sciatic nerves (n 3) were injected with CTB-conjugated rhodamine (10 mg/ml; List Biochemicals). CTB–rhodamine was also injected into normal sciatic nerve in control rats (n 2). The rats were allowed to survive for an additional two days before perfusion with Zamboni’s fixative, as described. The L4 and L5 DRGs after sciatic nerve injection were dissected for BDNF immunohistochemistry. Morphometric analysis The sizes of BDNF-IR neurons in DRGs of normal rats and of rats seven days after sciatic nerve lesion were measured by tracing circumferences of somata using the NIH Image program (1.59). At least 200 neurons were measured in each group. The different sizes of neurons
843
were allocated to different bins and plotted against the percentage of total neurons analysed. The size distribution of neurons labelled with both FluoroGold and BDNF, or BDNF only, was also analysed and frequency distribution histograms of cell size were plotted as described above. The percentage of neurons double labelled with both BDNF and CTB–rhodamine was analysed and compared between normal controls and after the nerve lesion. RESULTS
Effects of sciatic nerve lesion on the number of neurons immunoreactive for brain-derived neurotrophic factor in dorsal root ganglia In normal L4 and L5 DRGs, BDNF immunoreactivity was localized in a subpopulation of small and medium-sized sensory neurons (Fig. 1A, B). Low magnification of a DRG section showed that BDNF-IR neurons were distributed across all parts of the ganglia (Fig. 1A). High magnification showed that BDNF immunoreactivity is localized mainly in cytoplasm or axons close to somata, but not in nuclei (Fig. 1B). Some neurons were intensively labelled and some labelled only lightly. Cell counting data showed that 27 ^ 2% of sensory neurons were immunoreactive for BDNF (Table 1). This result is consistent with our previous findings. 57 One week after sciatic nerve lesion, the density and staining intensity of BDNF-IR neurons was significantly increased (Fig. 1C, D). High magnification showed that many BDNF-IR sensory axons were derived from, or associated with, largediameter neuronal somata (Fig. 1D). However, there was no significant change in BDNF staining in the contralateral DRG after the nerve lesion (Fig. 1E, F). To examine the time-course of changes in BDNF expression in the DRG after nerve lesion, the ipsiand contralateral DRGs were studied one, three, seven and 14 days after the nerve lesion. As shown in Fig. 2, BDNF staining began to increase 24 h after the nerve lesion (Fig. 2B, compared with the control in Fig. 2A). Both staining intensity and density of BDNF-IR neurons in the ipsilateral DRGs increased more significantly by three (Fig. 2C), seven (Fig. 2D) and 14 (Fig. 2E) days after the nerve lesion. There was no significant change in the contralateral DRG even two weeks after the nerve lesion (Fig. 2F). The quantitative data showed that the percentage of BDNF-IR neurons in the ipsilateral L5 DRG was increased to 33 ^ 2% (P , 0.05, compared with the control) 24 h after the lesion (Table 1). The increase in the percentage of BDNF-IR neurons in the L5 DRG was more obvious by three days (39%; P , 0.01 compared with the control and P , 0.05 compared with the contralateral side). By 14 days after the lesion, the percentage of BDNF-IR neurons increased to 44 ^ 4% (Fig. 1C, D; P , 0.01 compared with the control and P , 0.05 compared with the contralateral DRG). There was no
844
X.-F. Zhou et al.
Fig. 1. Low and high magnifications of DRG sections stained for BDNF. (A, B) Normal controls. (C, D) Ipsilateral DRG sections seven days after sciatic nerve lesion. (E, F) Contralateral DRG sections seven days after sciatic nerve lesion. Arrows in D indicate BDNF-IR axons. Scale bar in E also applies to A and C; scale bar in F also applies to B and D.
statistically significant difference at any time-point between the contralateral DRG and the normal DRG (Table 1). Effects of sciatic nerve lesion on the size distribution of brain-derived neurotrophic factor-immunoreactive sensory neurons In the control DRGs, BDNF-IR neurons were
small to medium in diameter, as shown in the size distribution histogram (Fig. 3A). A few largediameter neurons also contained BDNF immunoreactivity as demonstrated by a long tail towards large sizes. The peak of the size distribution curve fell at 400 mm 2. This result is consistent with our previous studies. Seven days after sciatic nerve lesion, many large-diameter neurons became BDNF-IR (Fig. 2D) and the size distribution
Injured sensory neurons switch phenotype for BDNF
845
Fig. 2. The time-course of BDNF expression in L5 DRGs after sciatic nerve lesion. (A–E) BDNF-IR neurons in ipsilateral DRGs 0, one, three, seven and 14 days after sciatic nerve lesion, respectively. (F) A contralateral DRG section three days postlesion. Arrows indicate axotomized sensory neurons with eccentric nuclei. Scale bar in F also applies to all other panels in this figure.
histogram shifted so that the peak fell at 1000 mm 2, a dramatic shift towards large sizes. As shown in Fig. 3A, the soma areas of 90% of the BDNF-IR neurons were smaller than 1000 mm 2 in the normal DRGs. In contrast, there were only 58% of neurons with soma areas less than 1000 mm 2 seven days after sciatic nerve lesion (P , 0.001 compared with the control, x 2 test). These results suggest that larger diameter neurons up-regulate BDNF expression after sciatic nerve lesion.
Switch of brain-derived neurotrophic factor phenotype in neurons after nerve lesion: demonstration by retrograde labelling Several questions arise from these findings. Firstly, do small neurons which normally express BDNF down-regulate BDNF synthesis after axotomy? Do all larger neurons up-regulate BDNF after axotomy? Since only 54% of all neurons in the L5 DRG project within the sciatic nerve, 42
846
X.-F. Zhou et al.
Table 1. Changes in percentages of brain-derived neurotrophic factor-immunoreactive neurons in L5 dorsal root ganglia after sciatic nerve lesion Postlesion time (days)
Ipsi (S.E.M.)
CL (S.E.M.)
0 1 3 7 14
26.6 ^ 2.3 32.6 ^ 1.8* 38.5 ^ 2.2**† 39.0 ^ 3.7** 44.0 ^ 3.8**†
28.8 ^ 1.5 29.0 ^ 3.4 31.2 ^ 2.7 30.5 ^ 3.0
*P , 0.05 compared with control; **P , 0.01 compared with control; †P , 0.05 compared with contralateral DRG. Values are means ^ S.E.M. (n 5). Ipsi, ipsilateral DRG; CL, contralateral DRG.
transection of the sciatic nerve results in axotomy of some, but not all, neurons in the L4 and L5 DRGs. Whether these small BDNF-IR neurons were intact was not possible to determine by single labelling. To address this question directly, all axotomized neurons were retrogradely labelled with FluoroGold and then double labelled with BDNF (Fig. 4A–D). The results showed that no axotomized small neurons with somata areas less than 600 mm 2 were immunoreactive for BDNF (Fig. 3B). When FluoroGold was injected into the intact sciatic nerve, it was easy to detect double-labelled, small-diameter neurons in DRGs (Fig. 4E, F), indicating that FluoroGold itself appears not to influence or mask the BDNF expression in small neurons. However, many large-diameter neurons were labelled by both FluoroGold and BDNF (Fig. 4A–D), indicating that axotomy resulted in an up-regulation of BDNF in large-diameter neurons. To further confirm this finding, we used CTB–rhodamine to label myelinated afferent neurons 50 and then double labelled with BDNF (Fig. 4G, H). Of the total CTB-labelled neurons (CTB–rhodamine injected into the lesioned sciatic nerve), 67 ^ 5% were BDNF-IR, which was significantly higher than in control rats (10 ^ 4%; CTB–rhodamine injected into intact sciatic nerves). Effects of sciatic nerve lesion on the expression of brain-derived neurotrophic factor messenger RNA in dorsal root ganglia Since BDNF protein is both anterogradely and retrogradely transported, BDNF immunoreactivity
Fig. 3. (A) Size distribution histogram for BDNF-IR neurons DRGs from control and sciatic nerve lesioned (Sci. n. Tx.) rats (seven days). Sciatic nerve lesion resulted in a dramatic shift of BDNF-IR neurons towards larger sizes. (B) Size distribution of injured (hatched bars) and intact (blank bars) BDNF-IR neurons in the L5 DRG seven days after sciatic nerve lesion.
in the DRG neurons may not necessarily represent the level of BDNF synthesis. To address this question, we studied BDNF mRNA expression in DRGs using RT–PCR and in situ hybridization techniques. In normal DRGs, BDNF mRNA is easily detectable after 26 cycles of amplification. The levels of BDNF mRNA increased as early as 16 h after nerve lesion (P , 0.05 compared with control and contralateral DRG; Fig. 5). This elevation was sustained and more marked with time after the lesion. There was no significant increase in the contralateral DRG at any time after nerve lesion (Fig. 5). As shown in Fig. 6A, a subpopulation of small neurons was labelled with an RNA probe complementary to BDNF mRNA (arrows). Some larger neurons were also faintly labelled, but no neuron was labelled if the sections were hybridized with
Fig. 4. Double labelling of sensory neurons in L4 and L5 DRGs for BDNF (B, D, F, H) and FluoroGold (A, C, E) or CTB–rhodamine (G) seven days after nerve lesion. (A, B) Arrows indicate intact neurons which were labelled for BDNF but not by FluoroGold applied to the transected sciatic nerve stump. Many large BDNF-IR neurons are double labelled with FluoroGold. (C, D) Higher magnification of micrographs from a lesioned DRG. Large arrowheads indicate large sensory neurons labelled with both BDNF and FluoroGold (note that these cells have eccentric nuclei); small arrowheads indicate small sensory neurons labelled with FluoroGold but not BDNF; asterisk indicates a large sensory neuron labelled with BDNF but not FluoroGold, indicating that it is an intact neuron (also indicated by the central position of its nucleus). (E, F) Micrographs from an intact DRG; arrows indicate small neurons labelled by both BDNF and FluoroGold, indicating that these neurons project to the sciatic nerve and BDNF expression is not influenced by FluoroGold. (G, H) Double-labelled neurons for CTB– rhodamine injected into the lesioned sciatic nerve (G) and BDNF (H). Arrow indicates a BDNF-labelled small neuron not labelled by CTB–rhodamine; arrowheads indicate double-labelled large sensory neurons. Scale bar in F also applies to A.
Injured sensory neurons switch phenotype for BDNF
Fig. 4.
847
848
X.-F. Zhou et al.
Fig. 5. Effects of sciatic nerve lesion on the levels of BDNF mRNA in DRGs as determined by an RT–PCR technique. The dagger and hatch signs indicate P , 0.05 compared with control and contalateral side, respectively. Gel panels under the histogram show PCR products from DRGs taken at different timepoints after sciatic nerve lesion. Lanes 1, 3, 5, 7 and 9 represent ipsilateral DRGs of control, 16 h, two, seven and 14 days after sciatic nerve lesion, respectively; lanes 2, 4, 6, 8 and 10 are from the corresponding contralateral DRGs. Ipsi, ipsilateral DRG; CL, contralateral DRG.
sense probe (data not shown). One day after sciatic nerve lesion, many larger sensory neurons were labelled with the BDNF antisense RNA probe (Fig. 6B). The reaction product seen around largediameter neurons (inset of Fig. 6B) may be from non-neuronal cells. The neurons surrounded by these glial cells were probably axotomized, as some of them had eccentric nuclei. There was no change in BDNF mRNA expression in the contralateral DRG. Two weeks after sciatic nerve lesion, more larger neurons expressed BDNF mRNA in the DRGs than in the control (Fig. 6C). BDNF mRNA signal was also found in Schwann cells in the DRG ipsilateral to the sciatic nerve transection. Effect of sciatic nerve lesion on brain-derived neurotrophic factor expression in the spinal cord
Fig. 6. Localization of BDNF mRNA in DRG sections after sciatic nerve lesion. (A) A control section of L5 DRG. (B) A DRG section 24 h after sciatic nerve lesion; inset shows that satellite cells around a large sensory neuron express BDNF mRNA. (C) A DRG section seven days after sciatic nerve lesion. Arrows in A indicate small neurons expressing BDNF mRNA; arrows in B indicate reaction product around large sensory neurons. Scale bar in C also applies to A and B.
In the control spinal cord (Fig. 7A), BDNF immunoreactivity was localized in nerve terminals in the superficial layers. A few BDNF nerve terminals were also detected in laminae III and IV, and the areas around the central canal. In the spinal cord 24 h after sciatic nerve lesion, BDNF-IR terminals increased in laminae III and IV (Fig. 7B) compared with control and contralateral dorsal horns (Fig. 7A). The increase in staining became more obvious three (Fig. 7C), seven (Fig. 7D) and 14 (Fig. 7E) days postlesion. Interestingly, the intensity of BDNF staining increased in the medial and lateral areas, but decreased in the central portion, of laminae I and II of the dorsal horn (see arrowhead in
Fig. 7D). The enlarged views of Fig. 7D showed that many BDNF-IR nerve terminals disappeared after the nerve lesion (Fig. 7G compared with Fig. 7H). These results indicate that BDNF expression is up-regulated in myelinated afferents, but downregulated in a subpopulation of unmyelinated afferents after axotomy. In line with our previous finding that BDNF is anterogradely transported in the intact unmyelinated afferents, 57 the increase in myelinated afferents after peripheral axotomy could be blocked by rhizotomy of the dorsal roots (see Fig. 7F).
Injured sensory neurons switch phenotype for BDNF
Fig. 7. Effects of sciatic nerve lesion on BDNF immunoreactivity in the spinal cord. (A) A normal control section of the spinal cord. (B) One day after sciatic nerve lesion. (C) Three days after the nerve lesion. (D) Seven days after the nerve lesion. (E) Two weeks after the nerve lesion. (F) Seven days after combined transection of the sciatic nerve and L4 and L5 dorsal roots. (G, H) High magnification of dorsal horns from the section in D; II in G, lamina II; arrow and arrowhead in G and H indicate the same location in D. Scale bar in A also applies to B–F; I and CL in B represent ipsilateral and contralateral sides, respectively, and also apply to other panels. IPSI in G, ipsilateral; CL in H, contralateral. Scale bar in H also applies to G.
849
850
X.-F. Zhou et al. DISCUSSION
This study has demonstrated that the expression of BDNF in DRGs after peripheral nerve injury is up-regulated at both the protein and mRNA levels. The time-course showed that the up-regulation occurs during the first 24 h and lasts for at least 14 days. Morphometric analysis and double labelling with retrograde tracers indicate that there is a cross-switch of BDNF phenotype between small and large sensory neurons. The differential distribution of BDNF also occurs in the spinal cord, where BDNF-IR fibres increase in deeper laminae but decrease in the central region of lamina II. These data suggest that BDNF may play an important role in the plasticity of sensory neurons after peripheral nerve injury.
Phenotypic switch A subpopulation of small to medium-sized sensory neurons in developing and adult animals expresses BDNF mRNA and protein. 14,57 In a previous study, Acheson et al. 1 showed that some isolated adult sensory neurons in vitro release BDNF, which affects the survival of these neurons in vitro. However, we have shown previously that BDNF synthesized by sensory neurons is anterogradely transported to their peripheral and central targets, and that a subpopulation of sensory neurons also transports BDNF retrogradely. 57 In the present study, we show that expression of BDNF in different subpopulations of sensory neurons is differentially regulated in response to axotomy. A subpopulation of small neurons which normally expresses BDNF down-regulates BDNF synthesis, whereas largediameter neurons up-regulate BDNF expression. Evidence from several experiments supports this conclusion. Firstly, many large-diameter neurons began to express BDNF 24 h after sciatic nerve lesion. A comparison of the size distribution of BDNF-IR neurons in normal and lesioned DRGs showed a dramatic shift from small to large sizes after nerve lesion. Secondly, BDNF immunohistochemistry combined with retrograde labelling of all axotomized sensory neurons showed that no axotomized neuron less than 600 mm 2 was BDNF-IR. No small sensory neurons labelled with BDNF contained FluoroGold, indicating that they were intact neurons not projecting in the sciatic nerve. Thirdly, most large sensory neurons immunoreactive for BDNF after sciatic nerve lesion were identified as having been axotomized by the presence of CTB–rhodamine, which was injected into the sciatic nerve, suggesting that BDNF is expressed in myelinated fibres. These results support the conclusion that subpopulations of sensory neurons cross-switch BDNF phenotype in response to axotomy.
Significance of the phenotypic switch Why do sensory neurons switch their expression of BDNF in response to peripheral nerve lesion? Recently, Neumann et al. 31 showed that sensory neurons switch SP gene expression in response to peripheral inflammation, which may result in allodynia due to an increased firing rate in the dorsal horn by stimulation of myelinated afferents. The phenotypic switch of sensory neurons immunoreactive for BDNF in response to peripheral nerve injury in this study may explain several well-described, but intriguing, observations reported over the past few years. For example, it is known that a subpopulation (about 30%) of mature sensory neurons can survive in the absence of neurotrophic factors in vitro, 1 as a result of an autocrine loop mediated by BDNF. Since there is also an increased number of sensory neurons in DRGs expressing TrkB following peripheral nerve injury, 16 it is likely that it is these neurons which survive in culture in the absence of additional neurotrophic factors. In contrast, small sensory neurons, which normally express BDNF, may undergo apoptosis due to down-regulation of BDNF or deprivation of target-derived NGF after peripheral nerve lesion. It is well known that a conditioning nerve lesion results in an enhanced regeneration of the same nerve, which cannot be prevented by application of antibodies for NGF to the cut end of the nerve. 37 In addition, previously axotomized sensory neurons cultured in vitro generate more and longer neurites than controls. 49 These studies suggest that factors intrinsic to sensory neurons play a role in regeneration. 10 Anterograde transport of BDNF expressed in sensory neurons is increased after nerve lesion, 47 so this phenomenon, together with an up-regulation of TrkB, 16 suggests that an autocrine mechanism operates in these neurons to enhance the nerve regeneration after a secondary nerve lesion. This mechanism may also explain the sprouting of myelinated afferents in the dorsal horn of the spinal cord in response to a sciatic nerve lesion. 50 Brain-derived neurotrophic factor is differentially regulated in axotomized neurons A number of studies suggests that BDNF synthesis in small neurons is regulated by NGF. Small neurons expressing BDNF also express the NGF high-affinity receptor, TrkA. 20 Injection of exogenous NGF results in a significant increase in BDNF expression. 3 NGF is synthesized in target tissues and retrogradely transported to a subpopulation of small to medium-sized neurons, 4,11 supports the survival of up to 70% of all sensory neurons during embryonic development, and regulates the synthesis of SP and CGRP in TrkA-containing neurons. 21,26 Disruption of sensory axons results in the reduction of SP and CGRP, 21 which can be prevented by introduction of exogenous NGF. 19,33,52 Small neurons which
Injured sensory neurons switch phenotype for BDNF
normally express BDNF switch off its expression after the nerve lesion, probably through an NGF– TrkA mechanism 28 in which injured small sensory neurons are deprived of NGF from their targets 3,8 or down-regulate NGF receptors. 34,55 Inflammation of the footpad, where NGF synthesis is increased, can result in an increase in the number and intensity of BDNF-IR neurons of the DRGs. 8 Thus, it is possible that down-regulation of BDNF in small neurons after nerve injury is due to the deprivation of sensory neurons of target-derived NGF. However, the expression of BDNF in largediameter neurons is normally suppressed and is not regulated by NGF, since these neurons do not contain NGF high-affinity receptor TrkA. It is likely that different BDNF gene promoters 46 are differentially involved in the regulation of BDNF expression in small and large sensory neurons after nerve lesion. 44 Glutamate and kainate are known to stimulate up-regulation of BDNF in many areas of the brain as a result of calcium influx. 45,48 Calcium influx triggers BDNF gene expression through a CREB transcription factor-dependent mechanism. 39,43 Based on the time-course of BDNF gene expression after sciatic nerve injury, which occurred as early as 16 h after injury, it is possible that direct or N-methyl-d-aspartate receptor-mediated calcium influx into neurons due to injury up-regulates BDNF gene in large-diameter neurons. The other possibility is that BDNF expression in large sensory neurons is controlled by NT-3, the survival factor for large sensory neurons. 13,15 A group of large
851
sensory neurons innervating Merkel cells depends on NT-3 for their survival 2 in postnatal mice and rats, but require BDNF for normal function of slowly adapting mechanotransduction. 6 It is known that the up-regulation of neuropeptide Y in large sensory neurons and in their central projections after the nerve lesion is diminished by application of exogenous NT-3 to the lesioned nerve. 32 We found that more than 96% of all neuropeptide Y-IR neurons in DRGs contained BDNF immunoreactivity after nerve lesion. 25 It is possible that normal BDNF expression in large sensory neurons is negatively regulated by target-derived NT-3. CONCLUSIONS
BDNF is differentially expressed in different subpopulations of sensory neurons in response to peripheral axotomy. Small neurons which normally express BDNF mRNA and protein down-regulate its expression, whereas large ones switch to a BDNF phenotype. The switch of BDNF phenotype in primary sensory neurons after nerve injury may explain changes in neuronal plasticity and may be involved in the induction of neuropathic pain after peripheral nerve injury.
Acknowledgements—We would like to thank Prof. E. Senba and Dr Ueyama from Wakayama Medical College, Japan for BDNF cDNA plasmids. This work was supported by NHMRC grants to X.F.Z. and R.A.R.
REFERENCES
1. Acheson A., Conover J. C., Fandl J. P., DeChlara T. M., Russell M., Thadani A., Squinto S. P., Yancopoulos G. D. and Lindsay R. M. (1995) A BDNF autocrine loop in adult sensory neurons prevents cell death. Nature 374, 450–453. 2. Airaksinen M. S., Koltzenburg M., Lewin G. R., Masu Y., Helbig C., Wolf E., Brem G., Toyka K. V., Thoenen H. and Meyer M. (1996) Specific subtypes of cutaneous mechanoreceptors require neurotrophin-3 following peripheral target innervation. Neuron 16, 287–295. 3. Apfel S. C., Wright D. E., Wiideman A. M., Dormia C., Snider W. D. and Kessler J. A. (1996) Nerve growth factor regulates the expression of brain-derived neurotrophic factor mRNA in the peripheral nervous system. Molec. cell. Neurosci. 7, 134–142. 4. Bandtlow C. E., Heumann R., Schwab M. E. and Thoenen H. (1987) Cellular localization of nerve growth factor synthesis by in situ hybridization. Eur. molec. Biol. Org. J. 6, 891–899. 5. Bennett D. L., French J., Priestley J. V. and McMahon S. B. (1996) NGF but not NT-3 or BDNF prevents the A fiber sprouting into lamina II of the spinal cord that occurs following axotomy. Molec. cell. Neurosci. 8, 211–220. 6. Carroll P., Lewin G. R., Koltzenburg M., Toyka V. and Thoenen H. (1998) A role for BDNF in mechanosensation. Nature Neurosci. 1, 42–46. 7. Chen C., Zhou X. F. and Rush R. A. (1996) Neurotrophin-3 and trkC-immunoreactive neurons in rat dorsal root ganglia correlate by distribution and morphology. Neurochem. Res. 21, 809–814. 8. Cho H. J., Kim J. K., Zhou X. F. and Rush R. A. (1997) Increased brain-derived neurotrophic factor immunoreactivity in rat dorsal root ganglia and spinal cord following peripheral inflammation. Brain Res. 764, 269–272. 9. Chomczynski P. and Sacchi N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Analyt. Biochem. 162, 156–159. 10. Chong M. S., Woolf C. J., Turmaine M., Emson P. C. and Anderson P. N. (1996) Intrinsic versus extrinsic factors in determining the regeneration of the central processes of rat dorsal root ganglion neurons—the influence of a peripheral nerve graft. J. comp. Neurol. 370, 97–104. 11. DiStefano P. S., Friedman B., Radziejewski C., Alexander C., Boland P., Schick C. M., Lindsay R. M. and Wiegand S. J. (1992) The neurotrophins BDNF, NT-3, and NGF display distinct patterns of retrograde axonal transport in peripheral and central neurons. Neuron 8, 983–993. 12. Ernfors P., Lee K.-F. and Jaenisch R. (1994) Mice lacking brain-derived neurotrophic factor develop with sensory deficits. Nature 368, 147–150. 13. Ernfors P., Lee K.-F., Kucera J. and Jaenisch R. (1994) Lack of neurotrophin-3 leads to deficiencies in the peripheral nervous system and loss of limb proprioceptive afferents. Cell 77, 503–512. 14. Ernfors P., Wetmore C., Olson L. and Persson H. (1990) Identification of cells in rat brain and peripheral tissues expressing mRNA for members of the nerve growth factor family. Neuron 5, 511–526.
852
X.-F. Zhou et al.
15. Farinas I., Jones K. R., Bachus C., Wang X.-Y. and Reichardt L. F. (1994) Severe sensory and sympathetic deficits in mice lacking neurotrophin-3. Nature 369, 658–661. 16. Foster E., Robertson B. and Fried K. (1994) TrkB-like immunoreactivity in rat dorsal root ganglia following sciatic nerve injury. Brain Res. 659, 267–271. 17. Hokfelt T., Zhang X. and Wiesenfeld-Hallin Z. (1994) Messenger plasticity in primary sensory neurons following axotomy and its functional implications. Trends Neurosci. 17, 22–30. 18. Janig W. and Schmidt R. F. (1992) In Reflex Sympathetic Dystrophy. Pathophysiological Mechanisms and Clinical Implications (eds Janig W. and Schmidt R. F.). VCH, Weinheim. 19. Jiang Z. G. and Smith R. A. (1995) Regulation by nerve growth factor of neuropeptide phenotypes in primary cultured sensory neurons prepared from aged as well as adult mice. Devl Brain Res. 90, 190–193. 20. Kashiba H., Ueda Y. and Senba E. (1996) Coexpression of preprotachykinin-a, alpha-calcitonin gene-related peptide, somatostatin, and neurotrophin receptor family messenger RNAs in rat dorsal root ganglion neurons. Neuroscience 70, 179–189. 21. Kessler J. A. and Black I. B. (1980) Nerve growth factor stimulates the development of substance P in sensory ganglia. Proc. natn. Acad. Sci. U.S.A. 77, 649–652. 22. Klein R., Silos-Santiago I., Smeyne R. J., Lira S. A., Brambilla R., Bryant S., Zhang L., Snider W. D. and Barbacid M. (1994) Disruption of the neurotrophin-3 receptor gene trkC eliminates Ia muscle afferents and results in abnormal movements. Nature 368, 249–251. 23. Levi-Montalcini R. (1987) The nerve growth factor: 35 years later. Science 237, 1154–1162. 24. Lewin G. R., Rueff A. and Mendell L. M. (1994) Peripheral and central mechanisms of NGF-induced hyperalgesia. Eur. J. Neurosci. 6, 1903–1912. 25. Li W. P., Xian C. J., Rush R. A. and Zhou X.-F. (1998) Upregulation of brain-derived neurotrophic factor and NPY in dorsal ascending sensory pathways in the rat. Neurosci. Lett. 260, 49–52. 26. Lindsay R. M. and Harmar A. J. (1989) Nerve growth factor regulates expression of neuropeptide genes in adult sensory neurons. Nature 337, 362–364. 27. McQuarrie I. G. and Grafstein B. (1973) Axonal outgrowth enhanced by previous nerve injury. Archs Neurol. 29, 53–55. 28. Michael G. J., Averill S., Nitkunan A., Rattray M., Bennett D. L., Yan Q. and Priestley J. V. (1997) Nerve growth factor treatment increases brain-derived neurotrophic factor selectively in TrkA-expressing dorsal root ganglion cells and in their central terminations within the spinal cord. J. Neurosci. 17, 8476–8490. 29. Molliver D. C., Wright D. E., Leitner M. L., Parsadanian A. S., Doster K., Wen D., Yan Q. and Snider W. D. (1997) IB4binding DRG neurons switch from NGF to GDNF dependence in early postnatal life. Neuron 19, 849–861. 30. Mu X., Silos-Santiago I., Carroll S. L. and Snider W. D. (1993) Neurotrophin receptor genes are expressed in distinct patterns in developing dorsal root ganglia. J. Neurosci. 13, 4029–4041. 31. Neumann S., Doubell T. P., Leslie T. and Woolf C. J. (1996) Inflammatory pain hypersensitivity mediated by phenotypic switch in myelinated primary sensory neurons. Nature 384, 360–364. 32. Ohara S., Tantuwaya V., DiStefano P. S. and Schmidt R. E. (1995) Exogenous NT-3 mitigates the transganglionic neuropeptide Y response to sciatic nerve injury. Brain Res. 699, 143–148. 33. Otten U., Lorez H. P. and Businger F. (1983) Nerve growth factor antagonizes the neurotoxic action of capsaicin on primary sensory neurones. Nature 301, 515–517. 34. Raivich G., Hellweg R. and Kreutzberg G. W. (1991) NGF receptor-mediated reduction in axonal NGF uptake and retrograde tranport following sciatic nerve injury and during regeneration. Neuron 7, 151–164. 35. Ramer M. and Bisby M. (1997) Reduced sympathetic sprouting occcurs in dorsal root ganglia after axotomy in mice lacking low affinity neurotrophin receptor. Neurosci. Lett. 228, 9–12. 36. Richardson P. M. and Issa V. M. K. (1984) Peripheral injury enhances central regeneration of primary sensory neurones. Nature 309, 791–793. 37. Richardson P. M. and Verge V. M. K. (1986) The induction of a regenerative propensity in sensory neurons following peripheral axonal injury. J. Neurocytol. 15, 585–594. 38. Ruit K. G., Elliott J. L., Osborne P. A., Yan Q. and Snider W. D. (1992) Selective dependence of mammalian dorsal root ganglion neurons on nerve growth factor during embryonic development. Neuron 8, 573–587. 39. Shieh P. B., Hu S. C., Bobb K., Timmusk T. and Ghosh A. (1998) Identification of a signaling pathway involved in calcium regulation of BDNF expression. Neuron 20, 727–740. 40. Smeyne R. J., Klein R., Schnapp A., Long L. K., Bryant S., Lewin A., Lira S. A. and Barbacid M. (1994) Severe sensory and sympathetic neuropathies in mice carrying a disrupted Trk/NGF receptor gene. Nature 368, 246–249. 41. Snider W. D. and McMahon S. B. (1998) Tackling pain at the source: new ideas about nociceptors. Neuron 20, 629–632. 42. Swett J. E., Torigoe Y., Elie V. R., Bourassa C. M. and Miller P. G. (1991) Sensory neurons of the rat sciatic nerve. Expl Neurol. 114, 82–103. 43. Tao X., Finkbeiner S., Arnold D. B., Shaywitz A. J. and Greenberg M. E. (1998) Ca 21 influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron 20, 709–726. 44. Timmusk T., Lendahl U., Funakoshi H., Arenas E., Persson H. and Metsis M. (1995) Identification of brain-derived neurotrophic factor promoter regions mediating tissue-specific, axotomy-, and neuronal activity-induced expression in transgenic mice. J. Cell Biol. 128, 185–199. 45. Timmusk T. and Metsis M. (1994) Regulation of BDNF promoters in the rat hippocampus. Neurochem. Int. 25, 11–15. 46. Timmusk T., Palm K., Metsis M., Reintam T., Paalme V., Saarma M. and Persson H. (1993) Multiple promoters direct tissuespecific expression of the rat BDNF gene. Neuron 10, 475–489. 47. Tonra J. R., Curtis R., Wong V., Cliffer K. D., Park J. S., Timmes A., Nguyen T., Lindsay R. M., Acheson A. and DiStefano P. S. (1998) Axotomy upregulates the anterograde transport and expression of brain-derived neurotrophic factor by sensory neurons. J. Neurosci. 18, 4374–4383. 48. Wetmore C., Olson L. and Bean A. J. (1994) Regulation of brain-derived neurotrophic factor (BDNF) expression and release from hippocampal neurons is mediated by non-NMDA type glutamate receptors. J. Neurosci. 14, 1688–1700. 49. White D. M., Mansfield K. and Kelleher K. (1996) Increased neurite outgrowth of cultured rat dorsal root ganglion cells following transection or inhibition of axonal transport of the sciatic nerve. Neurosci. Lett. 208, 93–96. 50. Woolf C. J., Shortland P. and Coggeshall R. E. (1992) Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature 355, 75–78.
Injured sensory neurons switch phenotype for BDNF
853
51. Woolf C. J., Shortland P., Reynolds M., Ridings J., Doubell T. and Coggeshall R. E. (1995) Reorganization of central terminals of myelinated primary afferents in the rat dorsal horn following peripheral axotomy. J. comp. Neurol. 360, 121–134. 52. Zhang Q., Ji R. R., Lindsay R. and Hokfelt T. (1995) Effect of growth factors on substance P mRNA expression in axotomized dorsal root ganglia. NeuroReport 6, 1309–1312. 53. Zhou X.-F., Deng Y.-S., Chie E. T., Xue Q., Zhong J.-H., McLachlan E. M., Rush R. A. and Xian C. (1999) Satellite cellderived neurotrophins induce noradrenergic sprouting in the dorsal root ganglia following peripheral nerve injury in the rat. Eur. J. Neurosci. (in press). 54. Zhou X.-F., Chie E. T. and Rush R. A. (1998) Distribution of brain-derived neurotrophic factor in cranial and spinal ganglia. Expl Neurol. 149, 237–242. 55. Zhou X.-F., Rush R. A. and McLachlan E. M. (1996) Differential expression of the p75 nerve growth factor receptor in glia and neurons of the rat dorsal root ganglia after peripheral nerve transection. J. Neurosci. 16, 2901–2911. 56. Zhou X.-F., Chie E. T., Deng Y. S. and Rush R. A. (1997) Mature sympathetic neurons derive neurotrophin-3 from effector tissues in rat. Eur. J. Neurosci. 9, 2753–2764. 57. Zhou X. F. and Rush R. A. (1996) Endogenous brain-derived neurotrophic factor is anterogradely transported in primary sensory neurons. Neuroscience 74, 945–951. (Accepted 8 January 1999)
Neuroscience Vol. 92, No. 3, pp. 841–853, 1999 Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00 S0306-4522(99)00027-5
INJURED PRIMARY SENSORY NEURONS SWITCH PHENOTYPE FOR BRAIN-DERIVED NEUROTROPHIC FACTOR IN THE RAT X.-F. ZHOU,*† E. T. CHIE,* Y.-S. DENG,* J.-H. ZHONG,* QING XUE,* R. A. RUSH* and C. J. XIAN‡ *Department of Human Physiology and Centre for Neuroscience, Flinders University of South Australia, GPO Box 2100, Adelaide 5001, Australia ‡Child Health Research Institute, North Adelaide 5006, Australia
Abstract—Peripheral nerve injury results in plastic changes in the dorsal root ganglia and spinal cord, and is often complicated with neuropathic pain. The mechanisms underlying these changes are not known. We have now investigated the expression of brain-derived neurotrophic factor in the dorsal root ganglia with histochemical and biochemical methods following sciatic nerve lesion in the rat. The percentage of neurons immunoreactive for brain-derived neurotrophic factor in the ipsilateral dorsal root ganglia was significantly increased as early as 24 h after the nerve lesion and the increase lasted for at least two weeks. The level of brain-derived neurotrophic factor messenger RNA was also significantly increased in the ipsibut not contralateral dorsal root ganglia. Both neurons and satellite cells in the lesioned dorsal root ganglia synthesized brain-derived neurotrophic factor messenger RNA after the nerve lesion. There was a dramatic shift in size distribution of positvie neurons towards large sizes seven days after sciatic nerve lesion. Morphometric analysis and retrograde tracing studies showed that no injured neurons smaller than 600 mm 2 were immunoreactive for brain-derived neurotrophic factor, whereas the majority of large injured neurons were immunoreactive in the ipsilateral dorsal root ganglia seven days postlesion. The brainderived neurotrophic factor-immunoreactive nerve terminals in the ipsilateral spinal cord were reduced in the central region of lamina II, but increased in more medial regions or deeper into laminae III/IV. These studies indicate that sciatic nerve injury results in a differential regulation of brain-derived neurotrophic factor in different subpopulations of sensory neurons in the dorsal root ganglia. Small neurons switched off their normal synthesis of brain-derived neurotrophic factor, whereas larger ones switched to a brain-derived neurotrophic factor phenotype. The phenotypic switch may have functional implications in neuronal plasticity and generation of neuropathic pain after nerve injury. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: BDNF, in situ hybridization, spinal cord, RT–PCR, sensory neurons.
Sciatic nerve injury results in distinct changes in sensory neurons of the dorsal root ganglia (DRGs) and their terminals in the spinal cord. In the adult, sciatic nerve lesion leads to a reduction in substance P (SP) and calcitonin gene-related peptide (CGRP), neurotransmitters for small nociceptive sensory neurons, and an increase in neuropeptide Y and vasoactive intestinal polypeptide in large sensory neurons. 17 Transection of the sciatic nerve triggers sprouting of myelinated sensory afferents from deep laminae into lamina II of the spinal cord, where the nociceptive fibres with high thresholds normally project. 49–51 A conditioning lesion of the sciatic nerve confers on sensory neurons an enhanced regenerative response to a secondary injury. 27,36 In †To whom correspondence should be addressed. Abbreviations: BDNF, brain-derived neurotrophic factor; CGRP, calcitonin gene-related peptide; CTB, cholera toxin B; DRG, dorsal root ganglion; GAPDH, glyceraldehyde-3phosphate dehydrogenase; IR, immunoreactive; NGF, nerve growth factor; NT-3, neurotrophin-3; PCR, polymerase chain reaction; RT, reverse transcription; SP, substance P; SSC, standard saline citrate. 841
addition to these morphological changes, nerve injury can induce clinical syndromes characterized by hyperalgesia and allodynia. 18 The mechanisms underlying these changes after nerve injury, however, are not known. Sensory neurons with unmyelinated or small-diameter myelinated fibres express TrkA and are dependent on nerve growth factor (NGF) for their survival. 23 Up to 70% of these neurons projecting to laminae I and II of the spinal cord are lost following gene deletion of TrkA or NGF, or upon administration of NGF antibody to developing animals. 38,40 These animals display hypoalgesia. 24 In contrast, large neurons expressing TrkC and containing neurotrophin-3 (NT-3) 7,30 are supported by NT-3 for their survival during development. 22,53 NT-3 null mutant mice lack proprioceptive and mechanoceptive functions, and the number of Merkel cells in touch domes is reduced. 2 Brain-derived neurotrophic factor (BDNF) and TrkB are normally synthesized in 20– 30% of all spinal sensory neurons and confined to those with small to medium diameters. 14,30 Deletion of the BDNF gene results in the reduction of sensory
842
X.-F. Zhou et al.
neurons in both spinal and cranial ganglia. The percentage of neuron loss varies between 30% in DRGs and 80% in vestibular ganglia. 12 BDNF is required for the normal function of slowly adapting mechanoceptive neurons, but not for their survival. 6 Although BDNF in sensory neurons is anterogradely transported to their peripheral and central targets, 57 it may also function as an autocrine neurotrophic factor. 1 A subpopulation of small nociceptive neurons which do not express any Trk receptors but express c-Ret and bind the lectin, IB4, depends on glial cell line-derived neurotrophic factor for survival. 29,41 It is likely that neurotrophic factors play important roles in neuronal plasticity and neuropathic pain after injury. Recently, we have found that satellite cell-derived NGF and NT-3 are involved in sympathetic sprouting in DRGs after nerve damage. 53 The low-affinity NGF receptor (p75) is also implicated in the formation of baskets of sympathetic nerves in DRGs. 35,55 NGF, but not BDNF or NT-3, delivered intrathecally, blocks the sprouting of myelinated afferents in the spinal cord after sciatic nerve lesion. 5 In a previous study, we found that about 25% of small to medium-sized sensory neurons in the normal rat lumbar DRGs are immunoreactive for BDNF, 57 which is consistent with the percentage of neurons containing mRNAs for BDNF. 14 Recently, it has been shown that peripheral nerve lesion results in an enhanced transport of both exogenous and endogenous BDNF in sensory neurons. 47 In the present study, we sought to determine what changes in BDNF expression occur in primary sensory neurons following sciatic nerve injury in rats, using a variety of techniques.
EXPERIMENTAL PROCEDURES
Animals Adult Sprague–Dawley rats, male and female (bred in Flinders University), were used under the guidelines of the National Health and Medical Research Council of Australia, and approved by the Animal Welfare Committee of Flinders University of South Australia. All efforts were made to minimize animal suffering and to reduce the number of animals.
Surgery Rats were anaesthetized with inhalational halothane, and the left sciatic nerves in the mid-thigh region were ligated and distally transected. Animals were allowed to survive for up to 14 days after lesion (n 5 for immunohistochemistry and n 3 for in situ hybridization for each time-point). Some rats without surgery were used as controls. Fresh L4 and L5 DRGs (n 3 or 4 for each time-point) were dissected bilaterally, frozen in liquid nitrogen and stored at 2708C prior to BDNF mRNA quantification. The remaining animals were perfused with Zamboni’s fixative and prepared for BDNF immunohistochemistry or BDNF in situ hybridization.
Immunohistochemistry of brain-derived neurotrophic factor BDNF immunohistochemistry was performed as described previously. The polyclonal rabbit antibody to BDNF made against recombinant human BDNF has been characterized previously and recognizes BDNF only, without cross-reactivity to other neurotrophins. 57 Rats were perfused with modified Zamboni’s fixative containing 2% formaldehyde and 15% saturated picric acid after an overdose of pentobarbital. After dissection, DRGs and the spinal cord were sectioned on a Cryostat microtome at 30 mm. Free-floating sections were treated with 0.3% hydrogen peroxide in 50% ethanol, blocked in 20% normal horse serum and incubated in rabbit polyclonal antibodies against recombinant human BDNF at a concentration of 1 mg/ml overnight. After extensive washing in phosphate-buffered saline (pH 7.4) containing 0.1% Tween 20, the sections were incubated in biotinylated secondary antibodies to rabbit immunoglobulin G, followed by incubation with the avidin–biotin–peroxidase complex reagent (ABC Kit, Vector Labs). Sections were developed in 0.05% diaminobenzidine containing 0.01% hydrogen peroxide and observed under a light microscope. All positive and negative neurons were counted from three randomly selected sections of L5 DRG using a Normaski optic setting. Only neurons with a visible nucleus were counted. The percentage of BDNF-immunoreactive (IR) neurons was then calculated for each animal, and the means from each group of animals were statistically compared using Student’s t-test.
Quantification of messenger RNA
brain-derived
neurotrophic
factor
The procedures for total RNA extraction and reverse transcription–polymerase chain reaction (RT–PCR) have been described previously. 54 In brief, total RNA was prepared from fresh rat ganglia (n 3) according to the acid guanidinium thiocyanate–phenol–chloroform method 9 using an RNA isolation kit (Advanced Biotechnologies, Leatherhead, U.K.). The RNA was treated with DNAase (Promega, U.S.A.) to remove possible contaminating genomic DNA and then subjected directly to first-strand cDNA synthesis by incubation with oligo dT15 and AMV reverse transcriptase. PCR primers for BDNF and GAPDH cDNAs were designed corresponding to the coding region of the genes as follows: BDNF primers, sense 5 0 TCCCTGGCTGACACTTTTGAG-3 0 and antisense 5 0 CTATCCTTATGAACCGCCAGC-3 0 ; GAPDH primers, sense 5 0 -TGCTGGTGCTGAGTATGTCG-3 0 and antisense 5 0 -GCATGTCAGATCCACAACGG-3 0 . PCR was performed in a 30-ml volume containing thermostable DNA polymerase (Advanced Biotechnologies, U.K.) on a Perkin DNA Thermal Cycler (Perkin Elmer, U.S.A.). All samples were heated at 958C for 2 min and amplified in cycles at 958C for 30 s, 588C for 30 s and 728C for 30 s. The last cycle was followed by a final incubation at 728C for 10 min. The house-keeping gene GAPDH was amplified in parallel to serve as an internal control. PCR cycle numbers for BDNF and GAPDH were optimized into the linear range. The appropriate concentration of cDNA was determined by a preliminary PCR and adjusted for quantitative PCR. The PCR products were electrophoresed and stained with ethidium bromide on a 1.5% agarose gel. The gel was captured as a digital image and analysed using a FluorImager 595 and quantitation software (Molecular Dynamics, U.S.A.). This analysis provided the same results as achieved with a previous method using radioactive hybridization. 56 The ratio of BDNF to GAPDH mRNAs was calculated from the fluorescent values of the PCR products. The mRNA level for each ganglion was expressed as a percentage of the mRNA level in the normal control ganglia.
Injured sensory neurons switch phenotype for BDNF
In situ hybridization In situ hybridization was used to localize BDNF mRNA expression in the tissues using digoxigenin-labelled riboprobes. BDNF antisense and sense riboprobes were made from a ribovector pGEM-3Zf(1) containing 800-bp rat BDNF after being linearized by EcoR I and Hind III and using SP6 and T7 RNA polymerase, respectively, using a digoxigenin-RNA labelling kit (Boehringer Mannheim, Mannheim, Germany). Paraffin sections of 4 mm were cut from tissue specimens, mounted on 3-aminopropyltriethoxysilane-coated glass slides, dewaxed, hydrated and treated with 0.2 N HCl for 20 min at room temperature. Sections were then permeabilized with 1 mg/ml proteinase K (Sigma) for 20 min at 378C. After postfixation for 5 min with 2% paraformaldehyde, sections were neutralized for 10 min with 0.2% glycine in phosphate-buffered saline. Following washing and dehydration, sections were covered with 25 ml hybridization mix (pre-heated for 5 min at 858C), containing 0.5 mg/ml sense (used as a negative control) or antisense probe, 50% formamide, 10% dextran sulphate, 0.05% Triton X-100, 500 mg/ml herring sperm DNA (Boehringer Mannheim), 0.05% polyvinylpyrrolidone and 5 × SSC (750 mM NaCl and 75 mM sodium citrate, pH 7.0). Sections were covered with glass coverslips and incubated in a humidified chamber for 18 h at 518C. Post-hybridization washes were undertaken at room temperature as follows with gentle shaking: once in 2 × SSC solution for 5 min to remove coverslips; twice in 2 × SSC for 15 min; and once in 1 × SSC for 10 min. After blocking of non-specific binding sites with 1% blocking reagent (Boehringer Mannheim) for 30 min, the sections were incubated for 30 min at 378C with an alkaline phosphatase-coupled sheep anti-digoxigenin immunoglobulin G (1:450; Boehringer Mannheim) and a colour reaction developed by incubating in nitroblue tetrazolium/X-phosphate substrate for 18 h in the dark according to the supplier’s instructions (Boehringer Mannheim). Sections were washed with water, counterstained lightly with 0.05% Methyl Green for 5 min and mounted with glycerol gel (Dako). Combination of retrograde labelling and brain-derived neurotrophic factor immunohistochemistry FluoroGold labelling. The left sciatic nerve in three rats was transected without ligation and the proximal stumps were soaked in 4% FluoroGold solution for 5 min. All lesioned neurons in L4 and L5 DRGs were labelled using this method. In control rats (n 2), FluoroGold (1 ml) was injected into the left sciatic nerve without lesion. The rats were allowed to survive for one week before BDNF immunohistochemistry on the DRGs. Cholera toxin B–rhodamine labelling. To label myelinated afferent neurons, cholera toxin B (CTB)–rhodamine tracer was used. After sciatic nerve lesion, the rats were reanaesthetized with inhalational halothane and the lesioned sciatic nerves (n 3) were injected with CTB-conjugated rhodamine (10 mg/ml; List Biochemicals). CTB–rhodamine was also injected into normal sciatic nerve in control rats (n 2). The rats were allowed to survive for an additional two days before perfusion with Zamboni’s fixative, as described. The L4 and L5 DRGs after sciatic nerve injection were dissected for BDNF immunohistochemistry. Morphometric analysis The sizes of BDNF-IR neurons in DRGs of normal rats and of rats seven days after sciatic nerve lesion were measured by tracing circumferences of somata using the NIH Image program (1.59). At least 200 neurons were measured in each group. The different sizes of neurons
843
were allocated to different bins and plotted against the percentage of total neurons analysed. The size distribution of neurons labelled with both FluoroGold and BDNF, or BDNF only, was also analysed and frequency distribution histograms of cell size were plotted as described above. The percentage of neurons double labelled with both BDNF and CTB–rhodamine was analysed and compared between normal controls and after the nerve lesion. RESULTS
Effects of sciatic nerve lesion on the number of neurons immunoreactive for brain-derived neurotrophic factor in dorsal root ganglia In normal L4 and L5 DRGs, BDNF immunoreactivity was localized in a subpopulation of small and medium-sized sensory neurons (Fig. 1A, B). Low magnification of a DRG section showed that BDNF-IR neurons were distributed across all parts of the ganglia (Fig. 1A). High magnification showed that BDNF immunoreactivity is localized mainly in cytoplasm or axons close to somata, but not in nuclei (Fig. 1B). Some neurons were intensively labelled and some labelled only lightly. Cell counting data showed that 27 ^ 2% of sensory neurons were immunoreactive for BDNF (Table 1). This result is consistent with our previous findings. 57 One week after sciatic nerve lesion, the density and staining intensity of BDNF-IR neurons was significantly increased (Fig. 1C, D). High magnification showed that many BDNF-IR sensory axons were derived from, or associated with, largediameter neuronal somata (Fig. 1D). However, there was no significant change in BDNF staining in the contralateral DRG after the nerve lesion (Fig. 1E, F). To examine the time-course of changes in BDNF expression in the DRG after nerve lesion, the ipsiand contralateral DRGs were studied one, three, seven and 14 days after the nerve lesion. As shown in Fig. 2, BDNF staining began to increase 24 h after the nerve lesion (Fig. 2B, compared with the control in Fig. 2A). Both staining intensity and density of BDNF-IR neurons in the ipsilateral DRGs increased more significantly by three (Fig. 2C), seven (Fig. 2D) and 14 (Fig. 2E) days after the nerve lesion. There was no significant change in the contralateral DRG even two weeks after the nerve lesion (Fig. 2F). The quantitative data showed that the percentage of BDNF-IR neurons in the ipsilateral L5 DRG was increased to 33 ^ 2% (P , 0.05, compared with the control) 24 h after the lesion (Table 1). The increase in the percentage of BDNF-IR neurons in the L5 DRG was more obvious by three days (39%; P , 0.01 compared with the control and P , 0.05 compared with the contralateral side). By 14 days after the lesion, the percentage of BDNF-IR neurons increased to 44 ^ 4% (Fig. 1C, D; P , 0.01 compared with the control and P , 0.05 compared with the contralateral DRG). There was no
844
X.-F. Zhou et al.
Fig. 1. Low and high magnifications of DRG sections stained for BDNF. (A, B) Normal controls. (C, D) Ipsilateral DRG sections seven days after sciatic nerve lesion. (E, F) Contralateral DRG sections seven days after sciatic nerve lesion. Arrows in D indicate BDNF-IR axons. Scale bar in E also applies to A and C; scale bar in F also applies to B and D.
statistically significant difference at any time-point between the contralateral DRG and the normal DRG (Table 1). Effects of sciatic nerve lesion on the size distribution of brain-derived neurotrophic factor-immunoreactive sensory neurons In the control DRGs, BDNF-IR neurons were
small to medium in diameter, as shown in the size distribution histogram (Fig. 3A). A few largediameter neurons also contained BDNF immunoreactivity as demonstrated by a long tail towards large sizes. The peak of the size distribution curve fell at 400 mm 2. This result is consistent with our previous studies. Seven days after sciatic nerve lesion, many large-diameter neurons became BDNF-IR (Fig. 2D) and the size distribution
Injured sensory neurons switch phenotype for BDNF
845
Fig. 2. The time-course of BDNF expression in L5 DRGs after sciatic nerve lesion. (A–E) BDNF-IR neurons in ipsilateral DRGs 0, one, three, seven and 14 days after sciatic nerve lesion, respectively. (F) A contralateral DRG section three days postlesion. Arrows indicate axotomized sensory neurons with eccentric nuclei. Scale bar in F also applies to all other panels in this figure.
histogram shifted so that the peak fell at 1000 mm 2, a dramatic shift towards large sizes. As shown in Fig. 3A, the soma areas of 90% of the BDNF-IR neurons were smaller than 1000 mm 2 in the normal DRGs. In contrast, there were only 58% of neurons with soma areas less than 1000 mm 2 seven days after sciatic nerve lesion (P , 0.001 compared with the control, x 2 test). These results suggest that larger diameter neurons up-regulate BDNF expression after sciatic nerve lesion.
Switch of brain-derived neurotrophic factor phenotype in neurons after nerve lesion: demonstration by retrograde labelling Several questions arise from these findings. Firstly, do small neurons which normally express BDNF down-regulate BDNF synthesis after axotomy? Do all larger neurons up-regulate BDNF after axotomy? Since only 54% of all neurons in the L5 DRG project within the sciatic nerve, 42
846
X.-F. Zhou et al.
Table 1. Changes in percentages of brain-derived neurotrophic factor-immunoreactive neurons in L5 dorsal root ganglia after sciatic nerve lesion Postlesion time (days)
Ipsi (S.E.M.)
CL (S.E.M.)
0 1 3 7 14
26.6 ^ 2.3 32.6 ^ 1.8* 38.5 ^ 2.2**† 39.0 ^ 3.7** 44.0 ^ 3.8**†
28.8 ^ 1.5 29.0 ^ 3.4 31.2 ^ 2.7 30.5 ^ 3.0
*P , 0.05 compared with control; **P , 0.01 compared with control; †P , 0.05 compared with contralateral DRG. Values are means ^ S.E.M. (n 5). Ipsi, ipsilateral DRG; CL, contralateral DRG.
transection of the sciatic nerve results in axotomy of some, but not all, neurons in the L4 and L5 DRGs. Whether these small BDNF-IR neurons were intact was not possible to determine by single labelling. To address this question directly, all axotomized neurons were retrogradely labelled with FluoroGold and then double labelled with BDNF (Fig. 4A–D). The results showed that no axotomized small neurons with somata areas less than 600 mm 2 were immunoreactive for BDNF (Fig. 3B). When FluoroGold was injected into the intact sciatic nerve, it was easy to detect double-labelled, small-diameter neurons in DRGs (Fig. 4E, F), indicating that FluoroGold itself appears not to influence or mask the BDNF expression in small neurons. However, many large-diameter neurons were labelled by both FluoroGold and BDNF (Fig. 4A–D), indicating that axotomy resulted in an up-regulation of BDNF in large-diameter neurons. To further confirm this finding, we used CTB–rhodamine to label myelinated afferent neurons 50 and then double labelled with BDNF (Fig. 4G, H). Of the total CTB-labelled neurons (CTB–rhodamine injected into the lesioned sciatic nerve), 67 ^ 5% were BDNF-IR, which was significantly higher than in control rats (10 ^ 4%; CTB–rhodamine injected into intact sciatic nerves). Effects of sciatic nerve lesion on the expression of brain-derived neurotrophic factor messenger RNA in dorsal root ganglia Since BDNF protein is both anterogradely and retrogradely transported, BDNF immunoreactivity
Fig. 3. (A) Size distribution histogram for BDNF-IR neurons DRGs from control and sciatic nerve lesioned (Sci. n. Tx.) rats (seven days). Sciatic nerve lesion resulted in a dramatic shift of BDNF-IR neurons towards larger sizes. (B) Size distribution of injured (hatched bars) and intact (blank bars) BDNF-IR neurons in the L5 DRG seven days after sciatic nerve lesion.
in the DRG neurons may not necessarily represent the level of BDNF synthesis. To address this question, we studied BDNF mRNA expression in DRGs using RT–PCR and in situ hybridization techniques. In normal DRGs, BDNF mRNA is easily detectable after 26 cycles of amplification. The levels of BDNF mRNA increased as early as 16 h after nerve lesion (P , 0.05 compared with control and contralateral DRG; Fig. 5). This elevation was sustained and more marked with time after the lesion. There was no significant increase in the contralateral DRG at any time after nerve lesion (Fig. 5). As shown in Fig. 6A, a subpopulation of small neurons was labelled with an RNA probe complementary to BDNF mRNA (arrows). Some larger neurons were also faintly labelled, but no neuron was labelled if the sections were hybridized with
Fig. 4. Double labelling of sensory neurons in L4 and L5 DRGs for BDNF (B, D, F, H) and FluoroGold (A, C, E) or CTB–rhodamine (G) seven days after nerve lesion. (A, B) Arrows indicate intact neurons which were labelled for BDNF but not by FluoroGold applied to the transected sciatic nerve stump. Many large BDNF-IR neurons are double labelled with FluoroGold. (C, D) Higher magnification of micrographs from a lesioned DRG. Large arrowheads indicate large sensory neurons labelled with both BDNF and FluoroGold (note that these cells have eccentric nuclei); small arrowheads indicate small sensory neurons labelled with FluoroGold but not BDNF; asterisk indicates a large sensory neuron labelled with BDNF but not FluoroGold, indicating that it is an intact neuron (also indicated by the central position of its nucleus). (E, F) Micrographs from an intact DRG; arrows indicate small neurons labelled by both BDNF and FluoroGold, indicating that these neurons project to the sciatic nerve and BDNF expression is not influenced by FluoroGold. (G, H) Double-labelled neurons for CTB– rhodamine injected into the lesioned sciatic nerve (G) and BDNF (H). Arrow indicates a BDNF-labelled small neuron not labelled by CTB–rhodamine; arrowheads indicate double-labelled large sensory neurons. Scale bar in F also applies to A.
Injured sensory neurons switch phenotype for BDNF
Fig. 4.
847
848
X.-F. Zhou et al.
Fig. 5. Effects of sciatic nerve lesion on the levels of BDNF mRNA in DRGs as determined by an RT–PCR technique. The dagger and hatch signs indicate P , 0.05 compared with control and contalateral side, respectively. Gel panels under the histogram show PCR products from DRGs taken at different timepoints after sciatic nerve lesion. Lanes 1, 3, 5, 7 and 9 represent ipsilateral DRGs of control, 16 h, two, seven and 14 days after sciatic nerve lesion, respectively; lanes 2, 4, 6, 8 and 10 are from the corresponding contralateral DRGs. Ipsi, ipsilateral DRG; CL, contralateral DRG.
sense probe (data not shown). One day after sciatic nerve lesion, many larger sensory neurons were labelled with the BDNF antisense RNA probe (Fig. 6B). The reaction product seen around largediameter neurons (inset of Fig. 6B) may be from non-neuronal cells. The neurons surrounded by these glial cells were probably axotomized, as some of them had eccentric nuclei. There was no change in BDNF mRNA expression in the contralateral DRG. Two weeks after sciatic nerve lesion, more larger neurons expressed BDNF mRNA in the DRGs than in the control (Fig. 6C). BDNF mRNA signal was also found in Schwann cells in the DRG ipsilateral to the sciatic nerve transection. Effect of sciatic nerve lesion on brain-derived neurotrophic factor expression in the spinal cord
Fig. 6. Localization of BDNF mRNA in DRG sections after sciatic nerve lesion. (A) A control section of L5 DRG. (B) A DRG section 24 h after sciatic nerve lesion; inset shows that satellite cells around a large sensory neuron express BDNF mRNA. (C) A DRG section seven days after sciatic nerve lesion. Arrows in A indicate small neurons expressing BDNF mRNA; arrows in B indicate reaction product around large sensory neurons. Scale bar in C also applies to A and B.
In the control spinal cord (Fig. 7A), BDNF immunoreactivity was localized in nerve terminals in the superficial layers. A few BDNF nerve terminals were also detected in laminae III and IV, and the areas around the central canal. In the spinal cord 24 h after sciatic nerve lesion, BDNF-IR terminals increased in laminae III and IV (Fig. 7B) compared with control and contralateral dorsal horns (Fig. 7A). The increase in staining became more obvious three (Fig. 7C), seven (Fig. 7D) and 14 (Fig. 7E) days postlesion. Interestingly, the intensity of BDNF staining increased in the medial and lateral areas, but decreased in the central portion, of laminae I and II of the dorsal horn (see arrowhead in
Fig. 7D). The enlarged views of Fig. 7D showed that many BDNF-IR nerve terminals disappeared after the nerve lesion (Fig. 7G compared with Fig. 7H). These results indicate that BDNF expression is up-regulated in myelinated afferents, but downregulated in a subpopulation of unmyelinated afferents after axotomy. In line with our previous finding that BDNF is anterogradely transported in the intact unmyelinated afferents, 57 the increase in myelinated afferents after peripheral axotomy could be blocked by rhizotomy of the dorsal roots (see Fig. 7F).
Injured sensory neurons switch phenotype for BDNF
Fig. 7. Effects of sciatic nerve lesion on BDNF immunoreactivity in the spinal cord. (A) A normal control section of the spinal cord. (B) One day after sciatic nerve lesion. (C) Three days after the nerve lesion. (D) Seven days after the nerve lesion. (E) Two weeks after the nerve lesion. (F) Seven days after combined transection of the sciatic nerve and L4 and L5 dorsal roots. (G, H) High magnification of dorsal horns from the section in D; II in G, lamina II; arrow and arrowhead in G and H indicate the same location in D. Scale bar in A also applies to B–F; I and CL in B represent ipsilateral and contralateral sides, respectively, and also apply to other panels. IPSI in G, ipsilateral; CL in H, contralateral. Scale bar in H also applies to G.
849
850
X.-F. Zhou et al. DISCUSSION
This study has demonstrated that the expression of BDNF in DRGs after peripheral nerve injury is up-regulated at both the protein and mRNA levels. The time-course showed that the up-regulation occurs during the first 24 h and lasts for at least 14 days. Morphometric analysis and double labelling with retrograde tracers indicate that there is a cross-switch of BDNF phenotype between small and large sensory neurons. The differential distribution of BDNF also occurs in the spinal cord, where BDNF-IR fibres increase in deeper laminae but decrease in the central region of lamina II. These data suggest that BDNF may play an important role in the plasticity of sensory neurons after peripheral nerve injury.
Phenotypic switch A subpopulation of small to medium-sized sensory neurons in developing and adult animals expresses BDNF mRNA and protein. 14,57 In a previous study, Acheson et al. 1 showed that some isolated adult sensory neurons in vitro release BDNF, which affects the survival of these neurons in vitro. However, we have shown previously that BDNF synthesized by sensory neurons is anterogradely transported to their peripheral and central targets, and that a subpopulation of sensory neurons also transports BDNF retrogradely. 57 In the present study, we show that expression of BDNF in different subpopulations of sensory neurons is differentially regulated in response to axotomy. A subpopulation of small neurons which normally expresses BDNF down-regulates BDNF synthesis, whereas largediameter neurons up-regulate BDNF expression. Evidence from several experiments supports this conclusion. Firstly, many large-diameter neurons began to express BDNF 24 h after sciatic nerve lesion. A comparison of the size distribution of BDNF-IR neurons in normal and lesioned DRGs showed a dramatic shift from small to large sizes after nerve lesion. Secondly, BDNF immunohistochemistry combined with retrograde labelling of all axotomized sensory neurons showed that no axotomized neuron less than 600 mm 2 was BDNF-IR. No small sensory neurons labelled with BDNF contained FluoroGold, indicating that they were intact neurons not projecting in the sciatic nerve. Thirdly, most large sensory neurons immunoreactive for BDNF after sciatic nerve lesion were identified as having been axotomized by the presence of CTB–rhodamine, which was injected into the sciatic nerve, suggesting that BDNF is expressed in myelinated fibres. These results support the conclusion that subpopulations of sensory neurons cross-switch BDNF phenotype in response to axotomy.
Significance of the phenotypic switch Why do sensory neurons switch their expression of BDNF in response to peripheral nerve lesion? Recently, Neumann et al. 31 showed that sensory neurons switch SP gene expression in response to peripheral inflammation, which may result in allodynia due to an increased firing rate in the dorsal horn by stimulation of myelinated afferents. The phenotypic switch of sensory neurons immunoreactive for BDNF in response to peripheral nerve injury in this study may explain several well-described, but intriguing, observations reported over the past few years. For example, it is known that a subpopulation (about 30%) of mature sensory neurons can survive in the absence of neurotrophic factors in vitro, 1 as a result of an autocrine loop mediated by BDNF. Since there is also an increased number of sensory neurons in DRGs expressing TrkB following peripheral nerve injury, 16 it is likely that it is these neurons which survive in culture in the absence of additional neurotrophic factors. In contrast, small sensory neurons, which normally express BDNF, may undergo apoptosis due to down-regulation of BDNF or deprivation of target-derived NGF after peripheral nerve lesion. It is well known that a conditioning nerve lesion results in an enhanced regeneration of the same nerve, which cannot be prevented by application of antibodies for NGF to the cut end of the nerve. 37 In addition, previously axotomized sensory neurons cultured in vitro generate more and longer neurites than controls. 49 These studies suggest that factors intrinsic to sensory neurons play a role in regeneration. 10 Anterograde transport of BDNF expressed in sensory neurons is increased after nerve lesion, 47 so this phenomenon, together with an up-regulation of TrkB, 16 suggests that an autocrine mechanism operates in these neurons to enhance the nerve regeneration after a secondary nerve lesion. This mechanism may also explain the sprouting of myelinated afferents in the dorsal horn of the spinal cord in response to a sciatic nerve lesion. 50 Brain-derived neurotrophic factor is differentially regulated in axotomized neurons A number of studies suggests that BDNF synthesis in small neurons is regulated by NGF. Small neurons expressing BDNF also express the NGF high-affinity receptor, TrkA. 20 Injection of exogenous NGF results in a significant increase in BDNF expression. 3 NGF is synthesized in target tissues and retrogradely transported to a subpopulation of small to medium-sized neurons, 4,11 supports the survival of up to 70% of all sensory neurons during embryonic development, and regulates the synthesis of SP and CGRP in TrkA-containing neurons. 21,26 Disruption of sensory axons results in the reduction of SP and CGRP, 21 which can be prevented by introduction of exogenous NGF. 19,33,52 Small neurons which
Injured sensory neurons switch phenotype for BDNF
normally express BDNF switch off its expression after the nerve lesion, probably through an NGF– TrkA mechanism 28 in which injured small sensory neurons are deprived of NGF from their targets 3,8 or down-regulate NGF receptors. 34,55 Inflammation of the footpad, where NGF synthesis is increased, can result in an increase in the number and intensity of BDNF-IR neurons of the DRGs. 8 Thus, it is possible that down-regulation of BDNF in small neurons after nerve injury is due to the deprivation of sensory neurons of target-derived NGF. However, the expression of BDNF in largediameter neurons is normally suppressed and is not regulated by NGF, since these neurons do not contain NGF high-affinity receptor TrkA. It is likely that different BDNF gene promoters 46 are differentially involved in the regulation of BDNF expression in small and large sensory neurons after nerve lesion. 44 Glutamate and kainate are known to stimulate up-regulation of BDNF in many areas of the brain as a result of calcium influx. 45,48 Calcium influx triggers BDNF gene expression through a CREB transcription factor-dependent mechanism. 39,43 Based on the time-course of BDNF gene expression after sciatic nerve injury, which occurred as early as 16 h after injury, it is possible that direct or N-methyl-d-aspartate receptor-mediated calcium influx into neurons due to injury up-regulates BDNF gene in large-diameter neurons. The other possibility is that BDNF expression in large sensory neurons is controlled by NT-3, the survival factor for large sensory neurons. 13,15 A group of large
851
sensory neurons innervating Merkel cells depends on NT-3 for their survival 2 in postnatal mice and rats, but require BDNF for normal function of slowly adapting mechanotransduction. 6 It is known that the up-regulation of neuropeptide Y in large sensory neurons and in their central projections after the nerve lesion is diminished by application of exogenous NT-3 to the lesioned nerve. 32 We found that more than 96% of all neuropeptide Y-IR neurons in DRGs contained BDNF immunoreactivity after nerve lesion. 25 It is possible that normal BDNF expression in large sensory neurons is negatively regulated by target-derived NT-3. CONCLUSIONS
BDNF is differentially expressed in different subpopulations of sensory neurons in response to peripheral axotomy. Small neurons which normally express BDNF mRNA and protein down-regulate its expression, whereas large ones switch to a BDNF phenotype. The switch of BDNF phenotype in primary sensory neurons after nerve injury may explain changes in neuronal plasticity and may be involved in the induction of neuropathic pain after peripheral nerve injury.
Acknowledgements—We would like to thank Prof. E. Senba and Dr Ueyama from Wakayama Medical College, Japan for BDNF cDNA plasmids. This work was supported by NHMRC grants to X.F.Z. and R.A.R.
REFERENCES
1. Acheson A., Conover J. C., Fandl J. P., DeChlara T. M., Russell M., Thadani A., Squinto S. P., Yancopoulos G. D. and Lindsay R. M. (1995) A BDNF autocrine loop in adult sensory neurons prevents cell death. Nature 374, 450–453. 2. Airaksinen M. S., Koltzenburg M., Lewin G. R., Masu Y., Helbig C., Wolf E., Brem G., Toyka K. V., Thoenen H. and Meyer M. (1996) Specific subtypes of cutaneous mechanoreceptors require neurotrophin-3 following peripheral target innervation. Neuron 16, 287–295. 3. Apfel S. C., Wright D. E., Wiideman A. M., Dormia C., Snider W. D. and Kessler J. A. (1996) Nerve growth factor regulates the expression of brain-derived neurotrophic factor mRNA in the peripheral nervous system. Molec. cell. Neurosci. 7, 134–142. 4. Bandtlow C. E., Heumann R., Schwab M. E. and Thoenen H. (1987) Cellular localization of nerve growth factor synthesis by in situ hybridization. Eur. molec. Biol. Org. J. 6, 891–899. 5. Bennett D. L., French J., Priestley J. V. and McMahon S. B. (1996) NGF but not NT-3 or BDNF prevents the A fiber sprouting into lamina II of the spinal cord that occurs following axotomy. Molec. cell. Neurosci. 8, 211–220. 6. Carroll P., Lewin G. R., Koltzenburg M., Toyka V. and Thoenen H. (1998) A role for BDNF in mechanosensation. Nature Neurosci. 1, 42–46. 7. Chen C., Zhou X. F. and Rush R. A. (1996) Neurotrophin-3 and trkC-immunoreactive neurons in rat dorsal root ganglia correlate by distribution and morphology. Neurochem. Res. 21, 809–814. 8. Cho H. J., Kim J. K., Zhou X. F. and Rush R. A. (1997) Increased brain-derived neurotrophic factor immunoreactivity in rat dorsal root ganglia and spinal cord following peripheral inflammation. Brain Res. 764, 269–272. 9. Chomczynski P. and Sacchi N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol–chloroform extraction. Analyt. Biochem. 162, 156–159. 10. Chong M. S., Woolf C. J., Turmaine M., Emson P. C. and Anderson P. N. (1996) Intrinsic versus extrinsic factors in determining the regeneration of the central processes of rat dorsal root ganglion neurons—the influence of a peripheral nerve graft. J. comp. Neurol. 370, 97–104. 11. DiStefano P. S., Friedman B., Radziejewski C., Alexander C., Boland P., Schick C. M., Lindsay R. M. and Wiegand S. J. (1992) The neurotrophins BDNF, NT-3, and NGF display distinct patterns of retrograde axonal transport in peripheral and central neurons. Neuron 8, 983–993. 12. Ernfors P., Lee K.-F. and Jaenisch R. (1994) Mice lacking brain-derived neurotrophic factor develop with sensory deficits. Nature 368, 147–150. 13. Ernfors P., Lee K.-F., Kucera J. and Jaenisch R. (1994) Lack of neurotrophin-3 leads to deficiencies in the peripheral nervous system and loss of limb proprioceptive afferents. Cell 77, 503–512. 14. Ernfors P., Wetmore C., Olson L. and Persson H. (1990) Identification of cells in rat brain and peripheral tissues expressing mRNA for members of the nerve growth factor family. Neuron 5, 511–526.
852
X.-F. Zhou et al.
15. Farinas I., Jones K. R., Bachus C., Wang X.-Y. and Reichardt L. F. (1994) Severe sensory and sympathetic deficits in mice lacking neurotrophin-3. Nature 369, 658–661. 16. Foster E., Robertson B. and Fried K. (1994) TrkB-like immunoreactivity in rat dorsal root ganglia following sciatic nerve injury. Brain Res. 659, 267–271. 17. Hokfelt T., Zhang X. and Wiesenfeld-Hallin Z. (1994) Messenger plasticity in primary sensory neurons following axotomy and its functional implications. Trends Neurosci. 17, 22–30. 18. Janig W. and Schmidt R. F. (1992) In Reflex Sympathetic Dystrophy. Pathophysiological Mechanisms and Clinical Implications (eds Janig W. and Schmidt R. F.). VCH, Weinheim. 19. Jiang Z. G. and Smith R. A. (1995) Regulation by nerve growth factor of neuropeptide phenotypes in primary cultured sensory neurons prepared from aged as well as adult mice. Devl Brain Res. 90, 190–193. 20. Kashiba H., Ueda Y. and Senba E. (1996) Coexpression of preprotachykinin-a, alpha-calcitonin gene-related peptide, somatostatin, and neurotrophin receptor family messenger RNAs in rat dorsal root ganglion neurons. Neuroscience 70, 179–189. 21. Kessler J. A. and Black I. B. (1980) Nerve growth factor stimulates the development of substance P in sensory ganglia. Proc. natn. Acad. Sci. U.S.A. 77, 649–652. 22. Klein R., Silos-Santiago I., Smeyne R. J., Lira S. A., Brambilla R., Bryant S., Zhang L., Snider W. D. and Barbacid M. (1994) Disruption of the neurotrophin-3 receptor gene trkC eliminates Ia muscle afferents and results in abnormal movements. Nature 368, 249–251. 23. Levi-Montalcini R. (1987) The nerve growth factor: 35 years later. Science 237, 1154–1162. 24. Lewin G. R., Rueff A. and Mendell L. M. (1994) Peripheral and central mechanisms of NGF-induced hyperalgesia. Eur. J. Neurosci. 6, 1903–1912. 25. Li W. P., Xian C. J., Rush R. A. and Zhou X.-F. (1998) Upregulation of brain-derived neurotrophic factor and NPY in dorsal ascending sensory pathways in the rat. Neurosci. Lett. 260, 49–52. 26. Lindsay R. M. and Harmar A. J. (1989) Nerve growth factor regulates expression of neuropeptide genes in adult sensory neurons. Nature 337, 362–364. 27. McQuarrie I. G. and Grafstein B. (1973) Axonal outgrowth enhanced by previous nerve injury. Archs Neurol. 29, 53–55. 28. Michael G. J., Averill S., Nitkunan A., Rattray M., Bennett D. L., Yan Q. and Priestley J. V. (1997) Nerve growth factor treatment increases brain-derived neurotrophic factor selectively in TrkA-expressing dorsal root ganglion cells and in their central terminations within the spinal cord. J. Neurosci. 17, 8476–8490. 29. Molliver D. C., Wright D. E., Leitner M. L., Parsadanian A. S., Doster K., Wen D., Yan Q. and Snider W. D. (1997) IB4binding DRG neurons switch from NGF to GDNF dependence in early postnatal life. Neuron 19, 849–861. 30. Mu X., Silos-Santiago I., Carroll S. L. and Snider W. D. (1993) Neurotrophin receptor genes are expressed in distinct patterns in developing dorsal root ganglia. J. Neurosci. 13, 4029–4041. 31. Neumann S., Doubell T. P., Leslie T. and Woolf C. J. (1996) Inflammatory pain hypersensitivity mediated by phenotypic switch in myelinated primary sensory neurons. Nature 384, 360–364. 32. Ohara S., Tantuwaya V., DiStefano P. S. and Schmidt R. E. (1995) Exogenous NT-3 mitigates the transganglionic neuropeptide Y response to sciatic nerve injury. Brain Res. 699, 143–148. 33. Otten U., Lorez H. P. and Businger F. (1983) Nerve growth factor antagonizes the neurotoxic action of capsaicin on primary sensory neurones. Nature 301, 515–517. 34. Raivich G., Hellweg R. and Kreutzberg G. W. (1991) NGF receptor-mediated reduction in axonal NGF uptake and retrograde tranport following sciatic nerve injury and during regeneration. Neuron 7, 151–164. 35. Ramer M. and Bisby M. (1997) Reduced sympathetic sprouting occcurs in dorsal root ganglia after axotomy in mice lacking low affinity neurotrophin receptor. Neurosci. Lett. 228, 9–12. 36. Richardson P. M. and Issa V. M. K. (1984) Peripheral injury enhances central regeneration of primary sensory neurones. Nature 309, 791–793. 37. Richardson P. M. and Verge V. M. K. (1986) The induction of a regenerative propensity in sensory neurons following peripheral axonal injury. J. Neurocytol. 15, 585–594. 38. Ruit K. G., Elliott J. L., Osborne P. A., Yan Q. and Snider W. D. (1992) Selective dependence of mammalian dorsal root ganglion neurons on nerve growth factor during embryonic development. Neuron 8, 573–587. 39. Shieh P. B., Hu S. C., Bobb K., Timmusk T. and Ghosh A. (1998) Identification of a signaling pathway involved in calcium regulation of BDNF expression. Neuron 20, 727–740. 40. Smeyne R. J., Klein R., Schnapp A., Long L. K., Bryant S., Lewin A., Lira S. A. and Barbacid M. (1994) Severe sensory and sympathetic neuropathies in mice carrying a disrupted Trk/NGF receptor gene. Nature 368, 246–249. 41. Snider W. D. and McMahon S. B. (1998) Tackling pain at the source: new ideas about nociceptors. Neuron 20, 629–632. 42. Swett J. E., Torigoe Y., Elie V. R., Bourassa C. M. and Miller P. G. (1991) Sensory neurons of the rat sciatic nerve. Expl Neurol. 114, 82–103. 43. Tao X., Finkbeiner S., Arnold D. B., Shaywitz A. J. and Greenberg M. E. (1998) Ca 21 influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron 20, 709–726. 44. Timmusk T., Lendahl U., Funakoshi H., Arenas E., Persson H. and Metsis M. (1995) Identification of brain-derived neurotrophic factor promoter regions mediating tissue-specific, axotomy-, and neuronal activity-induced expression in transgenic mice. J. Cell Biol. 128, 185–199. 45. Timmusk T. and Metsis M. (1994) Regulation of BDNF promoters in the rat hippocampus. Neurochem. Int. 25, 11–15. 46. Timmusk T., Palm K., Metsis M., Reintam T., Paalme V., Saarma M. and Persson H. (1993) Multiple promoters direct tissuespecific expression of the rat BDNF gene. Neuron 10, 475–489. 47. Tonra J. R., Curtis R., Wong V., Cliffer K. D., Park J. S., Timmes A., Nguyen T., Lindsay R. M., Acheson A. and DiStefano P. S. (1998) Axotomy upregulates the anterograde transport and expression of brain-derived neurotrophic factor by sensory neurons. J. Neurosci. 18, 4374–4383. 48. Wetmore C., Olson L. and Bean A. J. (1994) Regulation of brain-derived neurotrophic factor (BDNF) expression and release from hippocampal neurons is mediated by non-NMDA type glutamate receptors. J. Neurosci. 14, 1688–1700. 49. White D. M., Mansfield K. and Kelleher K. (1996) Increased neurite outgrowth of cultured rat dorsal root ganglion cells following transection or inhibition of axonal transport of the sciatic nerve. Neurosci. Lett. 208, 93–96. 50. Woolf C. J., Shortland P. and Coggeshall R. E. (1992) Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature 355, 75–78.
Injured sensory neurons switch phenotype for BDNF
853
51. Woolf C. J., Shortland P., Reynolds M., Ridings J., Doubell T. and Coggeshall R. E. (1995) Reorganization of central terminals of myelinated primary afferents in the rat dorsal horn following peripheral axotomy. J. comp. Neurol. 360, 121–134. 52. Zhang Q., Ji R. R., Lindsay R. and Hokfelt T. (1995) Effect of growth factors on substance P mRNA expression in axotomized dorsal root ganglia. NeuroReport 6, 1309–1312. 53. Zhou X.-F., Deng Y.-S., Chie E. T., Xue Q., Zhong J.-H., McLachlan E. M., Rush R. A. and Xian C. (1999) Satellite cellderived neurotrophins induce noradrenergic sprouting in the dorsal root ganglia following peripheral nerve injury in the rat. Eur. J. Neurosci. (in press). 54. Zhou X.-F., Chie E. T. and Rush R. A. (1998) Distribution of brain-derived neurotrophic factor in cranial and spinal ganglia. Expl Neurol. 149, 237–242. 55. Zhou X.-F., Rush R. A. and McLachlan E. M. (1996) Differential expression of the p75 nerve growth factor receptor in glia and neurons of the rat dorsal root ganglia after peripheral nerve transection. J. Neurosci. 16, 2901–2911. 56. Zhou X.-F., Chie E. T., Deng Y. S. and Rush R. A. (1997) Mature sympathetic neurons derive neurotrophin-3 from effector tissues in rat. Eur. J. Neurosci. 9, 2753–2764. 57. Zhou X. F. and Rush R. A. (1996) Endogenous brain-derived neurotrophic factor is anterogradely transported in primary sensory neurons. Neuroscience 74, 945–951. (Accepted 8 January 1999)