The evidence for nitric oxide synthase immunopositivity in the monosynaptic Ia-motoneuron pathway of the dog

Experimental Neurology 195 (2005) 161 – 178 www.elsevier.com/locate/yexnr

Regular Article

The evidence for nitric oxide synthase immunopositivity in the monosynaptic Ia-motoneuron pathway of the dog Jozef Marxalaa,*, Nade”da Luka´*ova´a, Igor Sˇullab, Peter Wohlfahrtc, Martin Marxalad Institute of Neurobiology, Slovak Academy of Sciences, Sˇolte´sovej 4, 040 01 Kosˇice, Slovak Republic b Department of Neurosurgery, Medical Faculty of P. J. Sˇafa´rik University, Kosˇice, Slovak Republic c Department of Histology and Embryology, Medical Faculty of P. J. Sˇafa´rik University, Kosˇice, Slovak Republic d Anesthesiology Research Laboratory, University of California, San Diego, La Jolla, CA 92037, USA a

Received 11 June 2004; revised 11 April 2005; accepted 20 April 2005 Available online 23 June 2005

Abstract In this study, nitric oxide synthase immunohistochemistry supported by nicotinamide adenine dinucleotide phosphate diaphorase histochemistry was used to demonstrate the nitric oxide synthase immunoreactivity in the monosynaptic Ia-motoneuron pathway exemplified by structural components of the afferent limb of the soleus H-reflex in the dog. A noticeable number of medium-sized intensely nitric oxide synthase immunoreactive somata (1000 – 2000 Am2 square area) and large intraganglionic nitric oxide synthase immunoreactive fibers, presumed to be Ia axons, was found in the L7 and S1 dorsal root ganglia. The existence of nitric oxide synthase immunoreactive fibers (6 – 8 Am in diameter, not counting the myelin sheath) was confirmed in L7 and S1 dorsal roots and in the medial bundle of both dorsal roots before entering the dorsal root entry zone. By virtue of the funicular organization of nitric oxide synthase immunoreactive fibers in the dorsal funiculus, the largest nitric oxide synthase immunoreactive fibers represent stem Ia axons located in the deep portion of the dorsal funiculus close to the dorsomedial margin of the dorsal horn. Upon entering the gray matter of L7 and S1 segments and passing through the medial half of the dorsal horn, tapered nitric oxide synthase immunoreactive collaterals of the stem Ia fibers pass through the deep layers of the dorsal horn and intermediate zone, and terminate in the group of homonymous motoneurons in L7 and S1 segments innervating the gastrocnemiussoleus muscles. Terminal fibers issued in the ventral horn intensely nitric oxide synthase immunoreactive terminals with long axis ranging from 0.7 to 15.1 Am presumed to be Ia bNOS-IR boutons. This finding is unique in that it focuses directly on nitric oxide synthase immunopositivity in the signalling transmitted by proprioceptive Ia fibers. Nitric oxide synthase immunoreactive boutons were found in the neuropil of Clarke_s column of L4 segment, varying greatly in size from 0.7 to 15.1 Am in length  0.7 to 4.8 Am wide. Subsequent to identification of the afferent nitric oxide synthase immunoreactive limb of the monosynaptic Ia-motoneuron pathway on control sections, intramuscular injections of the retrograde tracer Fluorogold into the gastrocnemius-soleus muscles, combined with nitric oxide synthase immunohistochemistry of L7 and S1 dorsal root ganglia, confirmed the existence of a number of medium-sized nitric oxide synthase immunoreactive somata (1000 – 2000 Am2 square area) in the dorsolateral part of both dorsal root ganglia, presumed to be proprioceptive Ia neurons. Concurrently, large nitric oxide synthase immunoreactive fibers were detected at the input and output side of both dorsal root ganglia. S1 and S2 dorsal rhizotomy caused a marked depletion of nitric oxide synthase immunoreactivity in the medial bundle of S1 and S2 dorsal roots and in the dorsal funiculus of S1, S2 and lower lumbar segments. In addition, anterograde degeneration of large nitric oxide synthase immunoreactive Ia fibers in the dorsal funiculus of L7-S2 segments produces direct evidence that the afferent limb of the soleus Hreflex is nitric oxide synthase immunoreactive and presents new immunohistochemical characteristics of the monosynaptic Ia-motoneuron pathway, unseparably coupled with the performance of the stretch reflex. D 2005 Elsevier Inc. All rights reserved. Keywords: Spinal cord; Dorsal root ganglia; Nitric oxide synthase; Monosynaptic Ia-pathway; Soleus H-reflex; Proprioceptive neurons

Introduction * Corresponding author. Fax: +421 55 6785 074. E-mail address: [email protected] (J. Marxala). 0014-4886/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.expneurol.2005.04.019

Immunohistochemical and histochemical detection of a morphologically heterogenous population of nitric oxide

162

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

synthase immunoreactive (bNOS-IR) and/or nicotinamide adenine dinucleotide phosphate diaphorase (NADPHd) exhibiting neurons, known to synthesize and release nitric oxide (NO) as a neuronal messenger and neurotransmitter (Bredt and Snyder, 1992; Garthwaite, 1991; Moncada and Higgs, 1993), has demonstrated that such neurons, even though small in number, are easily identifiable not only in various regions of the brain but also along the rostrocaudal axis of the spinal cord and dorsal root ganglia (DRGs) (Burnett et al., 1995; Luka´*ova´ et al., 1999; Marxala et al., 1997, 1998, 1999; Saito et al., 1994; Valtschanoff et al., 1992). It is thought that many, mostly small bNOS-IR and/or NADPHd-stained neurons in the DRGs, and especially those located in the superficial dorsal horn, may participate in the processing of nociceptive afferentation entering the spinal cord via the lateral bundle of the dorsal root(s). In line with this idea, a large number of studies has been performed on the role of NO in the transmission of nociceptive processing (Dykstra et al., 1993; Kitto et al., 1992; Luka´*ova´ et al., 2003, in press; Malmberg and Yaksh, 1993; Marxala et al., 2003; Meller and Gebhart, 1994; Meller et al., 1992a,b, 1993, 1994). In these studies different NO donors and inhibitors of nociceptive afferentation exhibited supporting as well as contradictory results (Aley et al., 1998; Babbedge et al., 1993; Kawabata et al., 1992, 1994; Sousa and Predo, 2001; Thompson et al., 1995; Wilson and Engbretson, 1998). The issue of whether bNOS and NO participate in the control of locomotion and reflex mechanisms, adapting the locomotor pattern to external demands transmitted by proprioceptive input including the monosynaptic Ia-motoneuron pathway during walking, remains unclear (Dietz, 2002). Nevertheless, previous in vivo studies have reported that 3-morpholinosydnonimine hydrochloride (SIN-1), an NO donor, potentiates electrically evoked monosynaptic reflexes in the cat spinal cord (Manjarrez et al., 2001); inhibition of bNOS activity in nervous tissue leads to decreased motor activity in the rat (Hal*a´k et al., 2000); and quite recently the enhancing role of NO on monosynaptic and polysynaptic spinal reflexes in cats has been noted (Tasci et al., 2003). In the search for the presence and origin of punctate neuronal nitric oxide synthase immunoreactivity in the neuropil of the phrenic nucleus in C3 – C5 segments, the occurrence and anterograde transport of neuronal NOS has been found in connection with the descending bulbospinal respiratory pathway terminating monosynaptically around the motoneurons of the phrenic nucleus (Marxala et al., 2002), and another recent study provides evidence for a hitherto unknown ascending premotor bNOS-IR pathway, monosynaptically connecting the neurons of the lumbosacral enlargement with the motoneurons of the ventral motor nucleus in the cervical enlargement of the dog (Marxala et al., 2004). Both findings allow us to hypothesize that neuronal NOS might be selectively localized in somata and

transported via long projecting axons which terminate monosynaptically, as shown in the case of the phrenic and ventral motor nucleus. The present study was intended to support this hypothesis by analyzing the structural components of the stretch reflex, choosing the soleus H-reflex as a prototypical monosynaptic reflex, with these objectives: (i) to test for the presence of neuronal NOS immunoreactivity in proprioceptive somata in L7 –S1 dorsal root ganglia (corresponding to the segmental distribution of homonymous motoneurons innervating the gastrocnemius medialis, lateralis and soleus muscles); (ii) to test for the presence of bNOS-IR in the peripheral and central processes of the proprioceptive somata; (iii) to ascertain the trajectories of bNOS-IR axons in the dorsal root entry zone and dorsal funiculus of L7 and S1 segments; and (iv) to identify the terminals of Ia bNOS-IR fibers and their target neurons in the gray matter of L7 and S1 segments, and in Clarke’s column of L4 segment.

Materials and methods Animal model, surgical procedures, tissue sampling, sectioning and staining of sections Adult mongrel dogs (n = 12) of both sexes weighing 9– 16 kg were used in these experiments. Experimental protocols were approved by the Institute of Neurobiology Animal Care Committee. The dogs were divided into three groups. In the first group of control animals (n = 5), the occurrence and distribution of bNOS-IR and/or NADPHdexhibiting somata and fibers in L7 and S1 DRGs, and the presence of bNOS-IR and/or NADPHd-stained fibers in the trunk of the tibial and sciatic nerve and L7 and S1 dorsal roots, were studied. In the same group the segments of the lumbosacral enlargement (L4– S2) in general, and L7 and S1 segments in particular were analyzed with emphasis on large-bNOS-IR fibers in the dorsal funiculus in both segments and their trajectory in the gray matter of both segments. In the second group (n = 3), Fluorogold (FG) injections into the gastrocnemius-soleus muscles were performed, and after 6 days of survival specimens from the trunk of the tibial and sciatic nerve were taken. L7 and S1 DRGs were dissected out bilaterally. Longitudinal serial sections (36 Am thick) were cut from both DRGs and processed alternately with FG-fluorescent microscopy and immunocytochemistry for bNOS-IR somata and fibers in L7 and S1 DRGs. In the third group (n = 4), unilateral dorsal rhizotomy of Cx1 and S3, S1 and S2, L6 and L7 and L4 and L5 dorsal roots midway between the DRG and the dorsal root entry zone was performed. The pairs of dorsal roots Cx1 and S3, S1 and S2, L6 and L7 and L4 and L5 were cut in different animals. After 6 days of survival, Cx1, S1 – S3 and L4 – L7

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

DRGs along with corresponding dorsal roots and lumbosacral enlargement (L4– S2 segments) were dissected out and processed for bNOS immunolabelling. Immunoreactivity of neuronal nitric oxide synthase and NADPH diaphorase staining of tibial and sciatic nerves, L7 and S1 DRGs, and L7 and S1 segments Immunoreactivity of bNOS was studied in a group of animals (n = 3) anesthetized with a mixture of ketamine and xylazine (100 and 15 mg/kg bw im) and artificially ventilated in a respirator with oxygen and nitrous oxide (Anemat N8 Chirana, CˇSSR). Afterwards the animals were perfused intracardially with heparinized saline and subsequently with freshly prepared 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS, pH 7.4). The specimens from the tibial and sciatic nerve were dissected out followed by a careful bilateral dissection of L7 and S1 DRGs and corresponding dorsal roots. The lumbosacral enlargement was removed in toto and postfixed in the same fixative for an additional 12 h. The following day the L4 – S2 segments were dissected and cryoprotected in PBS containing graded sucrose (15 –30%). The L4 – S2 segments were sectioned and at least 40 transverse serial sections (36 Am thick) from each segment were prepared. The specimens from the tibial and sciatic nerve were cut longitudinally. Thirty serial longitudinal sections (36 Am thick) were made from L7 and S1 DRGs. All sections were then processed for bNOS immunocytochemistry (Bredt et al., 1990). Free-floating tissue sections were pre-treated with 0.3% H2O2 in PBS for 30 min, then washed and blocked with 10% normal horse serum. For processing bNOS immunoreactivity, sections were incubated in Monoclonal AntiNitric Oxide Synthase-Brain antiserum (bNOS, Sigma, Product Number, N2280), 1:1000 dilution with 0.3% Triton X-100 in PBS overnight at 4-C with gentle agitation. This antibody specifically recognizes nitric oxide synthase (NOS) derived from rat brain (bNOS, 150 –160 kDa) and several breakdown products of lower molecular weight. It does not react with NOS derived from macrophages (mNOS) and endothelial cells (eNOS). To produce this antibody, a recombinant neuronal fragment (amino acids 1 – 181) from rat brain was used as immunogen (Dinerman et al., 1994). After several washes in PBS the sections were incubated with a biotinylated anti-mouse secondary antibody (1:200, Vector Laboratories) for 2 h. The next day the sections were rinsed with PBS and mixed with avidin– biotin complex (ABC) solution for 1 h (1:50; Vector Labs) at room temperature. Finally, the sections were developed in diaminobenzidine-H2O2 (DAB) as the chromogen, and on a few selected sections the bNOS staining was intensified by 1 min incubation in nickel chloride-enhancing solution. The sections were washed in distilled water, mounted on Colorfrost/Plus Microscope slides (Fisher Scientific, USA), air-dried and dehydrated

163

with graded alcohol (50 – 100%) followed by Xylene and coverslipped with Permount. On a few sections the primary antibody bNOS was omitted from the staining procedures so that no bNOS-positive cell bodies and fibers were detected. Two adult dogs, both males (n = 2) weighing 9 –16 kg, were used for NADPH diaphorase histochemistry. The animals were deeply anesthetized with pentobarbital (50 mg/kg iv) and perfused transcardially with saline followed by freshly prepared 4% paraformaldehyde + 0.1% glutaraldehyde buffered with 1 M sodium phosphate, pH 7.4. Following perfusion fixation the lumbosacral spinal cord and all lumbar and sacral DRGs were dissected out and stored in toto in the same fixative for 3– 4 h. Specimens from the tibial and sciatic nerve were taken and processed in the same way. After postfixation the lumbosacral spinal cord was divided into segments and L4 – S2 segments were cut in the transverse plane and at least 40 transverse serial sections (36 Am thick) were prepared from each segment. All lumbar and sacral DRGs were cut longitudinally. Similar sections were prepared from the tibial and sciatic nerve. All sections were processed for NADPHd activity using a modified histochemical procedure (Scherer-Singler et al., 1983) as follows: (1) The sections were incubated for 1 h at 37% in a solution containing 1 mg/ml nitroblue tetrazolium (NBT; Sigma, N-6876), 0.5 mg/ml of hNADPH (Sigma, N-1630), 0.8% Triton X-100 dissolved in 0.1 M phosphate buffer (pH 8.0) and 1.25 mg/ml monosodium malate (malic acid; Sigma, M-1125). (2) Control sections were treated in the same solution but without NADPH, thus testing for endogenous reduction activity in the corresponding blue formazan product. (3) The sections were then rinsed in 0.1 M phosphate buffer (pH 7.4), mounted on slides, air-dried overnight and coverslipped with Entellan. (4) Some sections were stained using the Nissl method to specify in more detail the laminar division of the spinal cord gray matter. Tracing of afferent pathways from the gastrocnemius-soleus muscles to the L7 and S1 DRGs using retrograde transport of Fluorogold (FG) Experiments were conducted on adult dogs (n = 3) weighing 9– 16 kg. The animals were anesthetized with a mixture of ketamine and xylazine (100 and 15 mg/kg bw im) and maintenance doses were delivered as necessary to maintain areflexia. For the intramuscular injections 25 – 30 Am of 4% solution of Fluorogold (FG; 2-hydroxy-4,4diamidinostilba midine; Biotium, Hayward, CA, USA) in 0.9 saline was injected under microscopic control through multiple penetrations (10 –12 2.5 Al aliquots) into the lateral and medial heads of the gastrocnemius muscle just before they fuse with each other distally and form a flat muscle. Muscle injections were performed unilaterally. After injections the needle was left in place for 2 –3 min to allow slow diffusion of the tracer. Sterile suture was used to close the

164

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

wounds. Animals recovered from the anesthesia within an hour and were returned to their cages to allow retrograde transport of the Fluorogold. Six days after the injections into the gastrocnemius-soleus muscles, the animals were reanesthetized with a mixture of ketamine and xylazine (as described above) and transcardially perfused with heparinized saline and subsequently with freshly prepared 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS; pH 7.4). Specimens from the tibial and sciatic nerve were dissected out and then the spinal cord and lumbosacral DRGs were exposed by dorsal laminectomy. The L7 and S1 DRGs (on the side exposed to FG) and the L7 and S1 spinal cord segments were removed and postfixed in the same fixative for an additional 12 h. The following day the L7 and S1 segments along with L7 and S1 DRGs were cryoprotected in PBS containing graded sucrose (15 – 30%). Thirty serial horizontally cut sections (36 Am thick) were prepared from both DRGs and one set of sections mounted on slides was used for FG-fluorescent microscopy, while alternate sections were collected separately in microdishes and used as free-floating tissue sections for bNOS immunocytochemistry (as described above). L7 and S1 segments were cut into 40 serial transverse sections, and FG-fluorescent microscopy was used to localize retrogradely FG-labelled motor neurons innervating the gastrocnemius-soleus muscles. Alternate sections from these series were processed for bNOS immunocytochemistry (Bredt et al., 1990). Dorsal rhizotomy-induced depletion of axonal nitric oxide synthase immunopositivity of fibers in the dorsal funiculus of the lumbosacral enlargement In the third group (n = 4), mid-lumbar (L4 and L5), lower lumbar (L6 and L7), sacral (S1 and S2) and sacrococcygeal (S3 and Cx1) dorsal roots were approached and divided unilaterally as follows. The animals were anesthetized with a mixture of ketamine and xylazine (100 and 15 mg/kg bw im) and artificially ventilated in a respirator with oxygen and nitrous oxide (Anemat N8 Chirana, CˇSSR). The L4 and L5, L6 and L7, S1 and S2 and S3 and Cxl dorsal roots were approached through lumbar laminectomy of the fifth to seventh laminas, thus gaining access to the cauda equina (Marxala et al., 1995; Orenda´*ova´ et al., 2000, 2001). The S3 and Cxl dorsal roots were divided 3– 4 mm proximally to the corresponding DRGs, whereas S1 and S2 dorsal rhizotomy was performed midway between S1 and S2 DRGs and the dural sac. A similar approach was adopted for L4 and L5, and L6 and L7 dorsal rhizotomy. Extreme care was taken to avoid damage to the radicular and spinal blood vessels. Following recovery from the anesthetic, each dog was transferred from the operating room to a holding cage for observation. Six days later the animals were deeply anesthetized (as above) and intracardially perfused with heparinized saline and sub-

sequently with freshly prepared 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS; pH 7.4). L4 – S2 segments of the spinal cord (i.e., the lumbosacral enlargement) and L4 – S2 DRGs were removed, postfixed and processed in turn in the same way as with the control material, including immunolabelling for bNOS. Counting bNOS-IR boutons in Clarke’s column of L4 segment and in laminas VII and IX of S1 segment For quantification of bNOS immunolabelled boutons, digital images of five sections taken from one side of Clarke’s column of L4 and laminas VII and IX of S1 spinal cord segments of control (n = 3) and both sides of dorsal rhizotomy-induced animals (n = 2) were captured using a light microscope (Olympus BX51), a digital camera (Olympus DP50) and Olympus DP-soft (version 3.0). The quantitative changes in bNOS labelling were determined by counting the number of bNOS-IR boutons, differing in length, within a frame of 25  25 Am in each section (altogether 5 sections/Clarke’s column of L4, 5 sections/ lamina VII of S1 and 5 sections/lamina IX of S1 per animal) using the UTHSCSA Image Tool 3. The size of bNOS-IR boutons was classified as follows: (1) 0.7 –3.5 Am in length  0.7 –2.8 Am wide; (2) 3.6 – 7.0 Am  2.9– 3.5 Am; (3) 7.1 – 15.0 Am  3.5– 4.1 Am; and (4)  15.1 Am  4.2– 4.8 Am. The results were statistically evaluated using ANOVA as well as the Tukey – Kramer test, and are expressed as means T SEM.

Results Identification and characteristics of bNOS-IR and/or NADPH diaphorase-exhibiting somata and fibers in L7 and S1 DRGs, and in the trunk of the tibial and sciatic nerve on control sections Immunolabelling for neuronal nitric oxide synthase revealed a noticeable proportion of nitric oxide synthesizing neurons in L7 and S1 DRGs (Figs. 1A and B). Although bNOS immunolabelling provided moderate or intense staining of multiple small (<1000 Am2/sq. a), medium-sized (1000 –2000 Am2/sq. a.) and some large (>2000 Am2/sq. a.) somata, little obvious topographic distribution of these small-, medium- and large-sized bNOS-IR cells could be detected (Figs. 2A, B, F, and G). However, a higher number of small- and medium-sized bNOS-IR somata were found in a slightly protruding dorsolateral portion of L7 and S1 DRGs opposite to the underlying motor root (Figs. 1A, B, 2B and G). A remarkable feature of the medium-sized bNOS-IR somata consisted in great inconsistency of immunolabelling, ranging from light immunolabelled cells more often seen in the central zone of DRGs (Fig. 2G) to intensely immunolabelled neurons with almost completely obscured

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

165

Fig. 1. Camera lucida drawing showing the distribution of small (<1000 Am2/sq. a., small dots), intensely and moderately labelled medium-sized (1000 – 2000 Am2/sq. a., medium-sized dots) and large (>2000 Am2/sq. a., large dots) bNOS-IR neurons in longitudinally cut left L7 (A) and left S1 (B) DRGs. Each dot represents one bNOS-IR soma; moderately labelled medium- or large-sized bNOS-IR neurons are shown as semicircles.

nuclei in the dorsolateral portion of both DRGs (Figs. 2C and H). Intraganglionic bNOS-IR fibers of various thickness including large ones (6– 8 Am in diameter) were identified just below the dorsolateral part of L7 and S1 DRGs (Figs. 2C, D, G, I, and J). Moreover, intense immunolabelling of the initial glomerulus could be detected in many medium-sized intensely immunolabelled bNOS-IR somata (Fig. 2E). A higher concentration of intensely NADPH diaphorase-stained fibers, 6 –8 Am in diameter (not counting the myelin sheath), was consistently found at the peripheral end of L7 and S1 DRGs (Fig. 3A). At the spinal end of L7 and S1 DRGs close to the exit of the central processes from both DRGs, evidently weaker fiber NADPH diaphorase staining was detected (Fig. 3B). A dense accumulation of bNOS-IR and/or NADPH diaphorase-stained fibers of varying thickness detected at the peripheral site of L7 and S1 DRGs was suggestive of the presence of such fibers in the lumbar and sacral nerves. After bNOS immunolabelling and/or for NADPHd staining of the tibial nerve (a more caudal terminal branch of the

sciatic nerve), it became apparent that the trunk of the tibial nerve giving off branches to the lateral and medial head of the gastrocnemius and soleus muscles appeared to contain a remarkably high number of bNOS-IR and/or NADPH diaphorase-stained axons (Fig. 3C). The majority of fibers were unmyelinated or thin myelinated axons. Nevertheless, a significant number of large myelinated axons 6 – 8 Am in diameter (not counting the myelin sheath), suggesting that they may represent large Ia fibers, were also NADPHdstained (Fig. 3D). Similar appearance and distribution of NADPHd-stained fibers was found in the trunk of the sciatic nerve (Fig. 3E). The distribution of nitric oxide synthase-immunoreactive axons in L7 and S1 dorsal roots and the organization of large nitric oxide synthase-immunoreactive axons in the dorsal funiculus of L7 and S1 segments on control sections The organization of large myelinated and small unmyelinated bNOS-IR fibers leaving L7 and S1 DRGs at their

166

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Fig. 2. Microphotographs depicting the distribution of bNOS-IR somata and fibers in L7 and S1 dorsal root ganglion. (A) A longitudinal section through L7 DRG; C—central end of DRG; P—peripheral end of DRG; arrow points to the dorsolateral part of DRG rich in bNOS-IR somata; arrowhead points to the ventral part of DRG which is poor in bNOS-IR somata. (B) The occurrence of small (S), medium-sized (M) and large (L) bNOS-IR somata in the dorsolateral part of L7 DRG; v—vessel. (C) A high-power microphotograph showing intensely immunolabelled medium-sized (M) and large (L) bNOS-IR somata; arrows point to bNOS-IR fibers in the dorsolateral part of DRG. (D) bNOS-IR fibers (asterisks) leaving the dorsolateral part of DRG. (E) Intensely immunolabelled bNOS-IR fibers (arrows) emerging from the medium-sized bNOS-IR somata (M). (F) A longitudinal section through S1 DRG; C—central end of DRG; P— peripheral end of DRG; arrows point to the dorsolateral part of DRG which is rich in bNOS-IR somata; arrowhead points to the ventral part of DRG which is poor in bNOS-IR somata. (G) A low-power microphotograph showing the dorsolateral (DL) and central (C) part of S1 DRG with many small and mediumsized bNOS-IR somata; bNOS-IR fibers are seen in the deep part of DRG (asterisks). (H) Medium-sized (M) and large (L) intensely immunolabelled somata in the dorsolateral part of S1 DRG. (I) Vertically (arrow) and horizontally (asterisk) oriented bNOS-IR fibers in the deep part of S1 DRG. (J) A higher-power magnification of large (L) and two medium-sized (M) intensely immunolabelled somata in the deep portion of S1 DRG; horizontally oriented bNOS-IR fibers (asterisks).

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

167

Fig. 3. (A) More intensely NADPHd-stained fibers (arrows) are seen at the peripheral end (P) of S1 DRG. Counterstained with neutral red. (B) Lightly NADPHd-stained fibers (arrows) seen at the central end (C) of S1 DRG; arrowhead points to an intensely NADPHd-stained soma. Counterstained with neutral red. (C) A longitudinal section through the tibial nerve; many large intensely (arrows) and a few fine (arrowhead) NADPHd-stained fibers are seen. (D) A highpower microphotograph showing large (arrows) NADPHd-stained fibers in the trunk of the tibial nerve (TN). (E) A microbundle of large (arrowheads) NADPHd-stained fibers in the trunk of the sciatic nerve (SN); oil immersion. (F) Large medial bundle (MB) of the S1 dorsal root containing large bNOS-IR fibers; DF—dorsal funiculus; DH—dorsal horn; LB—lateral bundle of the dorsal root. (G) Obliquely cut large bNOS-IR fibers (arrows) in the medial bundle. (H) Transversally cut large bNOS-IR fibers (arrowheads) of the medial bundle close to the dorsal horn (DH). (I) A low-power microphotograph of the dorsal funiculus (DF), dorsal horn (DH) and the dorsolateral funiculus (DLF). Fibers of the medial bundle (MB) of L7 dorsal root in the dorsal root entry zone. (J) Obliquely (asterisks) and transversally (arrowheads) cut large bNOS-IR fibers above the dorsal horn (DH). (K) An accumulation of large bNOS-IR fibers in the dorsomedial part of the dorsal funiculus (DF). Along the dorsal median septum (DMS) and in the triangle of Phillipe – Gombault (TPG), no large bNOS-IR fibers could be seen. (L) The occurrence of large, transversally cut bNOS-IR fibers (arrows) in the dorsal funiculus (DF) close to the medial border of the dorsal horn (DF).

168

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

spinal site and entering the corresponding dorsal root differs according to the rootlet position in the dorsal root. In the portion of the dorsal root between DRG and the pial ring, the small caliber unmyelinated bNOS-IR and/or NADPH diaphorase-stained fibers are loosely spread throughout the rootlet, and similarly large myelinated bNOS-IR axons appear uniformly distributed in the dorsal root. Approaching the dorsal root entry zone, about 1 mm before the pial ring, the small caliber bNOS-IR fibers move to the surface of the root and cross the pial ring. The large caliber bNOS-IR axons proceed centrally forming a massive bNOS-IR medial bundle of the dorsal root which enters the spinal cord forming a voluminous bundle between the dorsal funiculus and the dorsal horn (Figs. 3F – H). Once inside the dorsal funiculus large bNOS-IR fibers, many 6 – 8 Am in diameter, proceed more ventromedially and accumulate in the ventrolateral portion of the dorsal funiculus in close vicinity to the dorsomedial border of the upper half of the dorsal horn. The largest bNOS-IR fibers, before bifurcating into ascending and descending branches, tend to be located around the dorsomedial angle of the dorsal horn (Figs. 3I –L). In the dorsomedial and medial compartment of the dorsal funiculus in the segments studied, the density of such large bNOS-IR fibers is greatly reduced, and in the ventral-most portion of the dorsal funiculus the largest bNOS-IR axons are almost completely lacking. Tracing of afferent pathways from the gastrocnemius-soleus muscles to L7 and S1 DRGs using the retrograde transport of Fluorogold and immunoprocessing L7 and S1 DRGs for neuronal nitric oxide synthase The retrograde transport of FG injected unilaterally into the lateral and medial heads of the gastrocnemius and soleus muscles was used to trace afferents from these muscles to L7 and S1 DRGs ipsilateral with the injections. FG-labelled afferents arising in both heads of the gastrocnemius and soleus muscles could be detected in the trunk of the tibial nerve. Among the intensely FGfluorescent axons ranging between 1.5 and 2.5 Am in diameter identified on longitudinal sections, large FGfluorescent axonal profiles 6 –8 Am in diameter (Fig. 4A) were regularly seen. More proximally, FG-fluorescent axons were seen to proceed in the trunk of the sciatic nerve and enter L7 and S1 DRGs at their peripheral side. Most of the FG-labelled cell bodies, round or oval in shape with clearly distinguishable nuclei, were located in the dorsolateral part of L7 and S1 DRGs. It is noteworthy that the square area of most of the FGlabelled medium-sized cell bodies ranged between 1000 and 2000 Am2 (Figs. 4B –F). An interesting finding in both DRGs was the occurrence of FG-labelled fibers encircling large somata (Fig. 4G). Comparison of the spatial distribution of medium-sized FG-labelled neurons with intensely bNOS-IR somata of the same size and

location confirmed that a noticeable number of the FGfluorescent somata were immunoreactive for neuronal nitric oxide synthase (Figs. 4H –M). No FG-fluorescent fibers were found at the spinal side of L7 and S1 DRGs or in the corresponding dorsal roots. S1 and S2 dorsal rhizotomy and the loss of axonal bNOS immunopositivity in the dorsal root entry zone and dorsal funiculus of S1 and S2 segments Since the light microscopic appearance of the anterograde degeneration of bNOS-IR axons at the dorsal root entry zone and in the dorsal funiculus of four rhizotomized animals (i.e., Cx1 and S3, S1 and S2, L6 and L7 and L4 and L5) appeared to be very similar, only the results of S1 and S2 dorsal rhizotomy are given in more detail. Unilateral S1 and S2 dorsal rhizotomy followed by 6 days survival resulted in two readily identifiable phenomena: the depletion of axonal nitric oxide synthase immunopositivity and the disintegration of bNOS-IR axons in the dorsal funiculus of S1 and S2 segments ipsilateral with S1 and S2 rhizotomy. Dorsal rhizotomyinduced depletion of axonal bNOS immunopositivity played a key role in the sense that it removed bNOS, or more exactly it interrupted the anterograde axoplasmic transport of bNOS from presumed proprioceptive somata located in S1 and S2 DRGs through the medial bundle of S1 and S2 dorsal roots. Moreover, the depletion of axonal bNOS immunopositivity visualized as lightly stained fibers and background provided an outstanding contrast to the dark-brown axonal bNOS immunolabelling in the medial bundle of the non-axotomized dorsal root, dorsal root entry zone and in the dorsal funiculus contralateral to the dorsal rhizotomy (Figs. 5A –E). It is well established that the degeneration of a fiber system is a progressive process that eventually results in the complete disintegration of the entire system. In our experimental model it was found that for large presumed Ia bNOS-IR fibers a 6-day survival period after rhizotomy was long enough to allow for clear-cut degeneration of large bNOS-IR axons located in the dorsal funiculus in close vicinity to the dorsal and dorsomedial border of the dorsal horn in S1 and S2 segments (Figs. 5F and G). Transverse sections cut through degenerating large presumed Ia fibers demonstrated that axonal profiles and their axolemma lose their smooth circular outlines and tend to become wrinkled, distorted or beset with locally flattened expansions penetrating into the colorless myelin sheath (Figs. 5H – J). With the relatively short survival periods in our experimental model, the segmental extent of the fiber degeneration in the dorsal funiculus was relatively easy to determine, at least in sections cut through the lower and middle lumbar segments. The extent of the S1 and S2 dorsal rhizotomy-induced anterograde degeneration affecting both medial and lateral bundles of the corresponding

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

169

Fig. 4. (A) Arrows point to longitudinally cut FG-fluorescent axons in the trunk of the tibial nerve. P—perineurium. (B – F) FG—fluorescent medium-sized somata (arrows) in the dorsolateral part of L7 and S1 DRGs; CP—capsule. (G) FG—fluorescent fibers (arrows) encircling large somata in S1 DRG. (H) FG— fluorescent somata (FG-f) in L7 DRG after application of FG on the cut tibial nerve. Boxed areas in the dorsolateral and basal part of DRG containing FGfluorescent somata are enlarged in panels J and L. (I) A consecutive section cut through L7 DRG processed for bNOS immunoreactivity. Boxed areas in the dorsolateral and basal part of DRG containing bNOS-IR somata are enlarged in panels K and M. (J) A medium-sized FG-fluorescent cell body (arrow) in the dorsolateral part of L7 DRG. (K) A consecutive section processed for bNOS immunoreactivity; an identical bNOS-IR soma (arrow) is located in the dorsolateral part of DRG; CP—capsule. (L) A medium-sized FG-fluorescent cell body (arrow) in the basal part of DRG; CP—capsule. (M) A consecutive section processed for bNOS immunoreactivity; an identical bNOS-IR soma (arrow) is located in the basal part of DRG; CP—capsule.

dorsal roots changed according to segment level. While a high proportion of degenerating large caliber bNOS-IRmyelinated dorsal root axons of S1 and S2 dorsal root afferents entering the dorsal funiculus via the medial

bundle expanded into the dorsal funiculus after passing through the dorsal root entry zone, once there the degenerating axons covered a wide proportion of the dorsal funiculus. Proceeding rostrally a massive medial

170

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Fig. 5. (A) A low-power microphotograph showing depletion of axonal nitric oxide synthase immunopositivity and degeneration of bNOS-IR axons in the medial bundle of the dorsal root (MBRH) and in the dorsal funiculus (DFRH) of S1 segment after unilateral, right-sided S1 and S2 dorsal rhizotomy (RH). The left medial bundle (MB) and dorsal funiculus (DF) remained unchanged. DH—dorsal horn; v—vessel. (B – E) Normally appearing bNOS-IR fibers (arrows) in the medial bundle (MB) and dorsal funiculus (DF) on the non-rhizotomized side. (F – J) Depletion of axonal bNOS immunopositivity in the medial bundle of the dorsal root (MBRH) and dorsal funiculus (DFRH) and degeneration of large (Ia) bNOS-IR axons (arrows) in the dorsal funiculus on the rhizotomized side.

bundle of non-rhizotomized L7 dorsal root entered through the dorsal root entry zone, and degenerating S1 and S2 dorsal root fibers in the dorsal funiculus was gradually displaced more medially (Fig. 6). At the upper level of the lumbar enlargement in the L4 segment, the area of S1 and

S2 dorsal rhizotomy-induced axonal bNOS-IR degeneration was limited to a vertically oriented band of fibers in the dorsal funiculus located along the dorsal median septum and close to the dorsomedial angle of the dorsal funiculus.

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Fig. 6. A schematic depiction of the distribution of axonal degeneration in the dorsal funiculus of L7 – L5 segments after unilateral S2 and S1 dorsal rhizotomy. Moving cranially, the area of degenerating bNOS-IR axons shown as triangles shrinks and moves more medially, and, vice versa, undamaged bNOS-IR fibers of L6, L5, and L4 medial bundle indicated by black dots and bars of different size cover a large area in the dorsal funiculus. DFBP—dorsal funiculus, basal part.

The trajectory of bNOS-IR fibers and the appearance and distribution of bNOS-IR boutons in laminas VII and IX of L7 and S1 segments and in Clarke’s column of L4 segment, comparing sections from control animals with those after S1 and S2 dorsal rhizotomy The bNOS-IR collaterals appearing after bifurcation of the large bNOS-IR axons enter the dorsal horn from its dorsal and dorsomedial side, and before penetrating deeper into the gray matter they pass through the medial half of the dorsal horn. Passing through lamina VI most bNOS-IR fibers took a more ventrolateral or ventral course, while some fine bNOS-IR fibers are seen to ramify more ventrally, close to the interneurons. Most bNOS-IR fibers in the deep dorsal horn layers and the intermediate zone are mediumsized and thin fibers (Figs. 7A and B). Examination at high magnification of the neuropil in laminas VII and IX in L7 and S1 segments and in Clarke’s column of L4 segment demonstrated punctate bNOS staining, often having a bouton-like bNOS-IR appearance, juxtaposing some small unstained interneurons in lamina VII (Figs. 7C and D) or seen in close apposition to large unstained motoneurons in lamina IX of L7 and S1 segments and large dorsal spinocerebellar tract neurons in L4 segment (Fig. 7M). In control sections the bouton-like bNOS-IR structures – usually strongly bNOS immunolabelled – were most often oval in shape, or clublike, spherical or hemispherical, and sometimes elongated or crescentlike. An intriguing finding was the occurrence of bNOS-IR boutons (bNOS-IRBs) including giant ones at the base of

171

the dorsal horn (laminas V –VI), in laminas VII and IX of L7 and S1 segments and then in Clarke’s column of L4 segment. The long axis of bNOS-IRBs ranged from 0.7 to 22 Am. Giant bNOS-IRBs (above 7 Am) were found both in and outside Clarke’s column (L4), i.e., in laminas VII and IX of L7 and S1. The long axis of bNOS-IRBs sampled from the left lamina VII and lamina IX of L7 and S1 segment and Clarke’s column of L4 segment was used to categorize bNOS-IRBs into four subgroups as follows: (1) 0.7 – 3.5 Am in length  0.7 –2.8 Am in wide; (2) 3.6 – 7.0 Am  2.9 –3.5 Am; (3) 7.1 –15.0 Am  3.5– 4.1 Am; and (4) 15.1 Am  4.2 –4.8 Am (Figs. 7C – G, and I). Since large and giant bNOS-IR boutons may be similar in size to small neurons, special attention was given to the distribution of bNOS-IR interneurons in laminas VII and IX of L7 and S1 segments and in Clarke’s column of L4 segment. The identification of clearly bNOS-immunolabelled dendrites of interneurons in the corresponding laminas and segments on the one hand (Figs. 7S and T), and the occurrence of terminal bNOS-IR fibers attached to bNOS-IRBs on the other, can be taken as reliable criteria. Densitometric analysis enabling reliable detection of subtle differences in the optical density of bNOS immunostaining in the normal bouton-like structures demonstrates a prominent peak, seen as a narrow high triangle with the base regularly located in the left third of the densitogram and on the x-axis between 45 and 85 values of the 0- to 255-unit gray scale (Figs. 7H and J). Although the number of bNOS-IRBs of the first and second subgroups clearly prevails in laminas VII and IX of S1 segment, a frequency histogram of bNOS-IRBs shows that the number of bNOS-IRBs with the long axis above 15.1 Am clearly prevails (Fig. 8). S1 and S2 dorsal rhizotomy-induced changes in bNOSIRBs located in laminas VII and IX of S1, S2 and L7 segments and Clarke’s column of L4 segment ipsilateral with the rhizotomy are characterized by a boutonal depletion of bNOS immunopositivity, demonstrable mainly in the central part of many bNOS-IRBs (Figs. 7K, M, O, and Q). Considering the shape and extent of the densitograms taken from bNOS-IRBs in laminas VII and IX (L7) and Clarke’s column (L4) ipsilateral with the rhizotomy, it is apparent that the optical density of affected bNOS-IRBs is notably decreased and the broad base of the densitograms moves to the middle and right third of the x-axis (Figs. 7L, N, P, and R). Quantitative assessment of dorsal rhizotomyinduced changes in bNOS-IRBs shows that bNOS-IRBs are affected in all subgroups differentiated by bouton long axis measurement. Considering the shape and extent of the densitograms taken from bNOS-IRBs in laminas VII and IX of S1 segment and Clarke’s column of L4 segment, bNOS immunostaining seems to be substantially weaker in all rhizotomy-affected boutons. However, the boutonal bNOS-IR depletion is more apparent in the subgroups of bNOS-IRBs ranging from 7.1 Am to 15.1 Am in length (Fig. 9).

172

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Discussion The present study provides a description of immunohistochemically and histochemically identified structural components of the classic stretch reflex, exemplified by bNOS immunoreactivity and/or NADPH diaphorase staining in the following locations: in the afferent limb of the soleus Hreflex using the trunk of the tibial and sciatic nerve, the peripheral site of L7 and S1 DRGs, presumed proprioceptive (Ia) somata in L7 and S1 DRGs, the spinal site of L7 and S1 DRGs and L7 and S1 dorsal roots. Axonal bNOS immunoreactivity was also present in the dorsal funiculus of L7 and S1 segments, the lumbar enlargement (L4 –S2 segments) and finally in the course of terminal Ia bNOS-IR fibers and boutons including giant bNOS-IR boutons in the

gray matter of the lumbar enlargement in general, and in laminas VII and IX of L7 and S1 segments and Clarke’s column of L4 segment in particular. It follows from these observations that the whole afferent limb of the soleus H-reflex, arbitrarily divided into seven morphologically distinguishable subunits, can be identified using an immunohistochemical criterion as a presumed monosynaptic NOS-IR Ia pathway. In recent studies giving the functional characteristics of the stretch reflex including the structural components, i.e., the afferent and efferent limbs of the reflex arc (Crone and Nielsen, 1989; Dietz, 2002; Iles, 1986), and describing the active role of this reflex in the pathophysiology of spasticity (Crone et al., 1994; Delwaide and Olivier, 1988; Faist et al., 1994; Mazzocchio and Rossi, 1997; Morita et al., 2001;

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Nielsen et al., 1995), the presence of NOS in the afferent proprioceptive (Ia) neuron remained unnoticed. The reason why direct involvement of NOS in the monosynaptic Ia pathway was not considered consists perhaps in the lack of an immunohistochemical identification of NOS immunoreactivity in the proprioceptive neurons including somata in the corresponding DRGs and their peripheral fibers travelling proximally in the peripheral nerves, and through the dorsal roots to the spinal cord. It should be noted that controversial descriptions are given for peripheral fibers of proprioceptive neurons concerning their putative NOS immunoreactivity. While increased concentration of NOS was noted at the sarcolemma of muscle spindle fibers, in particular nuclear bag fibers, which belong among the type I fibers (Grozdanovic and Baumgarten, 1999), in normal rats sciatic nerve axons were NOS immunonegative (Gonza´lez-Herna´ndez and Rustioni, 1999a) and NADPH diaphorase staining, a histochemical marker for NOS, was virtually absent (Gonza´ lez-Herna´ ndez and Rustioni, 1999b). In contrast, the analysis of the control longitudinal sections taken from the trunk of the tibial and sciatic nerve in the dog and processed for bNOS and/or NADPH diaphorase histochemistry demonstrated clearly expressed axonal NOS immunoreactivity and/or NADPH diaphorasestained axons. Among them some large ones, 6 –8 Am in diameter and traceable for 500 – 800 Am, were easily recognized. The present study demonstrates that some medium-sized (1000 –2000 Am2/sq. a.) intensely bNOS-IR and/or heavily NADPH diaphorase-stained somata located mostly in the dorsolateral part of L7 and S1 DRGs of the dog may be proprioceptive neurons which with regard to the segmental level may represent the afferent bNOS-IR limb of the soleus H-reflex. While distinct populations of primary afferent neurons subserving different sensory modalities within various sensory ganglia have been identified using a variety of ultrastructural criteria and different histochemical

173

markers or immunohistochemistry (Dalsgaard et al., 1984; Gerebtzoff and Maeda, 1968; Ho¨kfelt et al., 1975; Knyiha´r, 1971; Molander et al., 1987; Nagy and Hunt, 1982; Rosenfeld et al., 1983; Sommer et al., 1985), the basic morphological and immunohistochemical characteristics of proprioceptive neurons in the DRGs are as yet not fully explained. NADPH diaphorase staining and peptide immunohistochemistry of rat DRGs revealed a peculiar pattern of NADPHd-exhibiting neurons in relation to spinal levels, and extremely low values were given for NADPHd-stained neurons in all cervical, upper thoracic, L2 – L4 and S2 – S3 DRGs (Aimi et al., 1991). In contrast to this, immunohistochemistry of NOS in the DRGs and spinal cord of man and rat revealed strong NOS immunolabelling in several primary sensory neurons, while the immunoreactive cells were of the small/medium-size type and were evenly distributed throughout the ganglia of all levels available, constituting approximately 40 – 50% of the whole cell population of the ganglia (Terenghi et al., 1993). The occurrence of medium-sized FG-fluorescent and bNOS-IR somata in the dorsolateral part of L7 and S1 DRGs speaks in favor of some topographic organization of these primary sensory cells (Kausz and Re´thelyi, 1985; Szarijanni and Re´thelyi, 1979). Analyzing the intraganglionic distribution and appearance of bNOS-IR and/or NADPHd-stained fibers in L7 and S1 DRGs, two interesting phenomena were noticed. First, an irregular network of bNOS-immunolabelled and/or NADPHd-exhibiting fibers was traceable through the central portion of both DRGs. Immunolabelled fibers of varying thickness ranging from very thin 0.8 –1.2 Am in diameter, often having a beaded appearance, to relatively large 6– 8 Am in diameter found just below the dorsolateral portion of DRGs. Second, noticeable differences were detected between caliber spectra of NOS-IR fibers at input and output, i.e., the spinal side of L7 and S1 DRGs. It is

Fig. 7. (A) A high-power microphotograph depicting a large bNOS-IR fiber (arrow) passing through the intermediate zone (LVII) in L7 segment in control material. (B) A high-power microphotograph showing a thin bNOS-IR fiber (arrowhead) passing through the lateral part of the intermediate zone (LVII) in L7 segment in control material. (C) A low-power microphotograph taken from the medial part of the intermediate zone (LVII) and showing bNOS-IR bouton (boxed area) in L7 segment in control material; LVIII—medial part of the ventral horn close to the ventral column (VC). (D) A high-power microphotograph showing a giant bNOS-IR bouton (gbt-boxed area in panel C) close to bNOS-unlabelled soma (asterisk). (E) A large bNOS-IR bouton (arrowhead) and terminal fiber (arrow) in the ventral horn (LIX) of S1 segment; control material. (F) An elongated giant bNOS-IR bouton (gbt) and terminal fiber (tf) in the ventral horn (LIX) of L7 segment; control material. (G) An intensely bNOS-immunolabelled giant bouton (gbt) and terminal fiber (tf) in Clarke’s column (C1c) of L4 segment: control material. (H) The densitogram of a giant bNOS-IR bouton (in panel G) is seen as a narrow high peak sharply rising and falling and the base located in the left third of x-axis. (I) A giant intensely immunolabelled bNOS bouton (gbt) and terminal fiber (tf) in Clarke’s column (C1c) of L4 segment (L4); DC—dorsal column; control material. (J) An extremely narrow peak of densitogram depicting optical density of giant bNOS-IR bouton (in panel I) in Clarke’s column of L4 segment. (K) A moderate bNOS depletion in a bouton (arrowhead) with terminal fiber (arrow) in lamina VII (LVII) of L7 segment ipsilateral with S1 and S2 dorsal rhizotomy (see details in text). (L) The base of the densitogram of the foregoing bNOS-IR bouton is located in the middle third of x-axis. (M) A giant bouton (gbt) apposing large dorsal spinocerebellar tract neuron (asterisk) in Clarke’s column (C1c) of L4 segment ipsilateral with S1 and S2 dorsal rhizotomy. Arrowhead points to the central, partly bNOS-depleted compartment of bouton. (N) The densitogram of the giant bNOS-IR bouton (in panel M) displaying two peaks and a broader base located more in the middle third of x-axis. The second peak corresponds with the central lightening of the bouton (arrowhead in panel M). (O) A giant strongly bNOS-IR depleted bouton (arrowhead—gbt) and terminal fiber (tf) in lamina IX (LIX) of L7 segment ipsilateral with S1 and S2 dorsal rhizotomy. (P) The densitogram of the giant strongly bNOS-depleted bouton (in panel O) with a broad base located in the middle and right third of x-axis and displaying several peaks. (Q) Similar giant strongly depleted bouton (arrowhead—gbt) and terminal fiber (tf) in the ventral horn (LIX) of L7 segment ipsilateral with S1 and S2 dorsal rhizotomy. (R) The densitogram of the foregoing giant bouton with the base located mostly in the middle third of x-axis. (S) Arrowhead points to an intensely bNOS-immunolabelled interneuron in the ventral horn (LIX) of L7 segment; d—dendrite; control material. (T) bNOS-IR interneuron (arrowhead) in Clarke’s column (C1c) of L4 segment; control material.

174

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Both dorsal roots break up into a series of rootlets 1– 2 Am before approaching the cord, and before entering the dorsal root entry zone (DREZ) the large and small bNOS-IR fibers segregate, the former creating a massive medial bundle in the L7 and S1 dorsal root which passes through the DREZ and enters the dorsolateral portion of the dorsal funiculus. Our

Fig. 8. The number of bNOS-IR boutons of four different subgroups (0.7 – 3.5 Am, 3.6 – 7.0 Am, 7.1 – 15.0 Am, 15.1 Am) seen on non-rhizotomized and rhizotomized side in lamina VII and lamina IX of S1 segment (A, B) and in Clarke’s column of L4 segment (C). Considering the shape and extent of the densitogram taken from bNOS-IR boutons in laminas VII, IX and Clarke’s column (Figs. 7H, J, L, N, P, and R), bNOS immunostaining seems to be substantially weaker in all rhizotomy-affected boutons and appeared to be more intensely stained for bNOS-IR in boutons on nonrhizotomized side. The number of bNOS-IR boutons (7.1 – 15.0 Am, 15.1 Am) was higher in lamina VII and in lamina IX of S1 segment (A, B) and in Clarke’s column of L4 segment (C) on rhizotomized side of the spinal cord. However, the number of bNOS-IR boutons ranging from 0.7 to 7.0 Am was not changed by dorsal rhizotomy.

known from classic studies that the central processes of DRG cells are finer than the peripheral processes (Ramo´n y Cajal, 1952), but this finding was not confirmed for large cells with myelinated processes (Lieberman, 1976). In relatively long L7 and S1 dorsal roots, the large bNOSIR axons appear uniformly distributed in the root, and there is no preferential localization of axons according to their size.

Fig. 9. The number of normally appearing and axotomy-induced degeneration of bNOS-IRBs in four different subgroups, differing in long axis (0.7 – 3.5 Am, 3.6 – 7.0 Am, 7.1 – 15.0 Am, 15.1 Am) in lamina VII and lamina IX of S1 segment (A, B) and in Clarke’s column of L4 segment (C) of the spinal cord after S1 and S2 dorsal rhizotomy and 6 days survival. Considering the shape and extent of the densitograms taken from bNOSIRBs in laminas VII, IX and Clarke’s column (see Figs. 7L, N, P, and R), bNOS immunostaining seems to be substantially weaker in all rhizotomyaffected boutons; this change is more apparent in the subgroups of bNOSIRBs ranging from 7.1 Am to 15.1 Am.

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

results are fully comparable with the anatomical distinction between large and small caliber fibers, albeit with no signs of the bNOS immunoreactivity given for the spinal cord-rootlet junction of the cat (Szenta´gothai, 1958), and some parceling of coarse and fine axons as in the monkey proximal rootlets, but not in the cat (Snyder, 1977). As can be understood from our findings, significant additions to the original descriptions of the primary afferent fibers (Ishizuka et al., 1979; Scheibel and Scheibel, 1969) consist in the identification and characteristics of bNOS-IR fibers, which taken topographically overlap completely with the primary afferent fibers that are known to be divided during partial selective dorsal rhizotomy aimed at alleviating spasticity without noticeable sensory deterioration (Fasano et al., 1976; Laitinen and Fugl-Meyer, 1982; Laitinen et al., 1983; Sindou et al., 1974). In order to compare the effect of dorsal rhizotomy on the monosynaptic Ia bNOS-IR pathway with that seen in humans after lumbar and sacral dorsal rhizotomy or dorsal root entry zone-otomy (DREZotomy; Lazorthes et al., 2002), decreasing monosynaptic and polysynaptic reflexes thus relieving spasticity, several coccygeal, sacral and lumbar dorsal roots were divided in our study. Processing corresponding DRGs and spinal cord segments for bNOS immunoreactivity after unilateral S1 and S2 dorsal rhizotomy, a triple effect of dorsal root division was found. First, a strongly enhanced retrograde axotomy-induced reaction of small and mainly medium-sized bNOS-IR somata in both DRGs was noted, consisting in dense immunolabelling of cytoplasm and intraganglionic fibers which appeared to be clearly detectable also at the peripheral and spinal side of DRGs. The intensity of dorsal rhizotomy-induced bNOS immunolabelling was comparable with the retrograde response of bNOS-IR neurons occurring in the ventral and dorsal respiratory group of the medulla after high cervical hemisection in the dog (Marxala et al., 2002), or in the bNOS-IR somata located in the lumbosacral segments projecting axons via the premotor ascending bNOS-IR pathway to motor neurons of the ventral motor nucleus in the cervical intumescence (Marxala et al., 2004). Second, while anterograde axotomy-induced degeneration of S1 and S2 dorsal roots was readily detectable in the medial and lateral band of both dorsal roots, the degeneration in the former was marked by strong depletion of axonal bNOS immunopositivity. Third, anterograde degeneration of the large bNOS-IR axons 6 – 8 Am in diameter (not counting the myelin sheath) aggregating in the deep portion of the dorsal funiculus along the dorsal and dorsomedial border of the dorsal horn in S1 and S2 segments suggests that they may in fact be stem Ia axons (Ishizuka et al., 1979) forming an intramedullar portion of the monosynaptic Ia pathway before bifurcating into ascending and descending branches in the dorsal funiculus. The presence of neuronal NOS immunoreactivity in Ia fibers forming the afferent limb of the soleus H-reflex, traceable from the trunk of the tibial nerve to bNOS-

175

immunoreactive boutons in the ventral horn (lamina IX) and intermediate zone (lamina VII) of the L7 and S1 segments, and in close vicinity with large neurons of Clarke’s column in the L4 segment, implies that bNOS may be directly involved in the afferent limb of the soleus H-reflex arc, and therefore together with l-glutamate (Maxwell et al., 1990) it can influence the transmission at monosynaptic Ia synapses. A number of previous light- and electron-microscopic studies on the synaptic connections between identified primary muscle afferent and motoneurons, interneurons and dorsal spinocerebellar tract neurons have revealed that these boutons vary greatly in size, from 1  1 Am to ‘‘giant’’ boutons of 20  3 Am, and while certain of these boutons would appear to fit the category ‘‘giant boutons’’, no clear criteria for defining giant boutons has so far been presented (Hongo et al., 1987). Originally the size of giant boutons was given as 10 – 15  6 – 8 Am in Clarke’s column (Szenta´gothai and Albert, 1955) and later it was proposed that boutons over 5 Am in length be regarded as giant (Kuno et al., 1973; Saito, 1974), although much larger boutons, 8– 10 Am long and occasionally 4 –5 Am wide in Clarke’s column of the cat, were found at electronmicroscopic level (Re´thelyi, 1970), and the sizes of en passant and terminal boutons were found to vary over a considerable range (1  1 to 20  3 Am) in Clarke’s column of the cat at lightmicroscopic level (Tracey and Walmsley, 1984). Based on observations of Ia primary afferent terminations in laminas V – VII and in motor nuclei (lamina IX) in L7 and S1 segments (Brown and Fyffe, 1978, 1981; Fyffe and Light, 1984) and in and outside of Clarke’s column in L3 and L4 segments of the cat, it was proposed that only boutons greater than 7 Am in length be regarded as giant boutons (Hongo et al., 1987). In the present study frequency histograms of the long axis (0.7 to 15.1 Am) of bNOSIR boutons sampled from lamina VII and laminas IX of L7 and S1 segments and Clarke’s column of L4 segment were prepared, and these boutons were divided into four subgroups considering their long axis. In line with Hongo’s classification (Hongo et al., 1987) bNOS-IR boutons of the third (7.1 – 15.0 Am in length) and fourth (15.1 Am) subgroup found in laminas VII and IX of L7 and S1 segments and in Clarke’s column of L4 segment belong to the category of giant boutons. Although no experimental data can be given fully explaining the size differences between the length of the synaptic contacts reported in the ventral horn of the lumbosacral enlargement (Brown and Fyffe, 1981; Conradi et al., 1983; Fyffe and Light, 1984; Pierce and Mendell, 1993) and those in our study, there are at least three causative factors which should be taken into account. First, the dorsal funiculi of L7 and S1 segments (along with L6 and S2) are rich in extremely large bNOS-IR axons. Second, it should be admitted that the size of animals used in the present study (dogs weighing 9 –16 kg) is hardly similar to the size of cats used in previous studies (Hongo et al., 1987, cats 2.2 –3.0 kg; Walmsley et al., 1985, cats 1.5 to 2.5 kg; Nicol and Walmsley, 1991, cats 1.5 to 3.5 kg; Pierce

176

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

and Mendell, 1993, young adult cats). Third, the thickness of the terminal part of bNOS-IR axonal arbor and the size of the branch issuing terminal may influence the size of the bouton (Fig. 19 in Hongo et al., 1987). At electronmicroscopic level the typical bouton in the ventral horn of the lumbar enlargement of the cat had an ovoid shape, varying from 0.7 to 7.0 Am along the long axis, and from 0.7 to 2.8 Am along the short axis. Bouton volumes ranged widely, from 0.45 to 22.4 Am3, but the bouton volume was related to the position within the afferent arbor (Pierce and Mendell, 1993). The presence of bNOS immunoreactivity in the large dorsal root afferents, dorsal root entry zone and dorsal funiculus, confirmed by dorsal rhizotomy-induced axonal and boutonal depletion of bNOS immunolabelling in the dorsal funiculus and terminal fields of Ia afferents, suggests that bNOS and nitric oxide may participate in anterograde signalling in the way described in the monosynaptic bulbospinal respiratory nitric oxide synthase-immunoreactive pathway and in the monosynaptic premotor nitric oxide synthase-immunoreactive pathway connecting lumbar segments with the ventral motor nucleus of the cervical enlargement in the dog (Marxala et al., 2002, 2004). It is reasonable to suggest that the presence of nitric oxide synthase and hence NO at the presynaptic site of Ia synapse can be an important factor influencing spinal reflexes. Recent studies strongly support this view, showing that NO donor SIN-I potentiates monosynaptic reflexes in the cat spinal cord (Manjarrez et al., 2001), and another arguing that systemic doses of sodium nitroprusside (SNP), an NO donor, seem more effective than local doses in enhancing the amplitude of the mono- and polysynaptic reflexes in anesthetized and spinalized cats (Tasci et al., 2003). Finally, the occurrence of neuronal nitric oxide synthase immunopositivity all along the afferent limb of the monosynaptic Ia pathway may be related directly with bNOS activity, decreased levels of which may reduce motor activity (Hal*a´k et al., 2000), or inhibit walking speed in rats (Wang et al., 2001).

Acknowledgments The authors thank Mr. D. Krokavec, Ms. M. Sˇponta´kova´, Mrs. M. Synekova´, Mrs. M. Vargova´ and Mrs. A.M. Koxova´ for their excellent technical assistance. The experimental work was supported by the VEGA Grants No. 2/3217/23 and 2/5134/25, from the SAS, APVT Grant No. 51-013002 and by NIH grants NS 32794 and NS 40386 to M.M.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.expneurol.2005.04.019.

References Aimi, Y., Fujimura, M., Vincent, S.R., Kimura, H., 1991. Localization of NADPH-diaphorase-containing neurons in sensory ganglia of the rat. J. Comp. Neurol. 306, 382 – 392. Aley, K.O., McCarter, G., Levine, J.D., 1998. Nitric oxide signaling in pain and nociceptor sensitization in the rat. J. Neurosci. 18, 7008 – 7014. Babbedge, R.C., Hart, S.L., Moore, P.K., 1993. Antinociceptive activity of nitric oxide synthase inhibitors in the mouse: dissociation between the effect of l-NAME an l-NMMA. J. Pharm. Pharmacol. 45, 77 – 79. Bredt, D.S., Snyder, S.H., 1992. Nitric oxide, a novel neuronal messenger. Neuron 8, 3 – 11. Bredt, D.S., Hwang, P.M., Snyder, S.H., 1990. Localization of nitric oxide synthase indicating a neural role for nitric oxide. Nature 347, 768 – 770. Brown, A.G., Fyffe, R.E., 1978. The morphology of group Ia afferent fibre collaterals in the spinal cord of the cat. J. Physiol. 274, 111 – 127. Brown, A.G., Fyffe, R.E., 1981. Direct observations on the contacts made between Ia afferent fibres and alpha-motoneurones in the cat’s lumbosacral spinal cord. J. Physiol. 313, 121 – 140. Burnett, A.L., Saito, S., Maguire, M.P., Yamaguchi, H., Chang, T.S., Hanley, D.F., 1995. Localization of nitric oxide synthase in spinal nuclei innervating pelvic ganglia. J. Urol. 153, 212 – 217. Conradi, S., Cullheim, S., Gollvik, L., Kellerth, J.-O., 1983. Electron microscopic observation on the synaptic contacts of group Ia muscle spinal afferents in the cat lumbosacral spinal cord. Brain Res. 265, 31 – 39. Crone, C., Nielsen, J., 1989. Methodological implications of the post activation depression of the soleus H-reflex in man. Exp. Brain Res. 78, 28 – 32. Crone, C., Nielsen, J., Petersen, N., Ballegaard, M., Hultborn, H., 1994. Disynaptic reciprocal inhibition of ankle extensors in spastic patients. Brain 117, 1161 – 1168. Dalsgaard, C.-L., Ygge, J., Vincent, S.R., Ohrling, M., Dockray, G.J., Elde, R., 1984. Peripheral projections and neuropeptide coexistence in a subpopulation of fluoride-resistant acid phosphatase reactive spinal primary sensory neurons. Neurosci. Lett. 51, 139 – 144. Delwaide, P.J., Olivier, E., 1988. Short-latency autogenic inhibition (IB inhibition) in human spasticity. J. Neurol., Neurosurg. Psychiatry 51, 1546 – 1550. Dietz, V., 2002. Proprioception and locomotor disorders. Nat. Rev., Neurosci. 3, 781 – 790. Dinerman, J.L., Dawson, T.M., Schell, M.J., Snowman, A., Snyder, S.H., 1994. Endothelial nitric oxide synthase localized to hippocampal pyramidal cells: implications for synaptic plasticity. Proc. Natl. Acad. Sci. U. S. A. 91, 4214 – 4218. Dykstra, C., Meller, S.T., Gebhart, G.F., 1993. Mechanisms of thermal hyperalgesia: receptor subtypes and intracellular events. IASP, 224. Faist, M., Mazevet, D., Dietz, V., Pierrot-Deseilligny, E., 1994. A quantitative assessment of presynaptic inhibition of Ia afferents in spastics. Differences in hemiplegics and paraplegics. Brain 117, 1449 – 1455. Fasano, V.A., Barolat-Romana, G., Ivaldi, A., Sguazzi, A., 1976. Functional posterior radiculotomy, in the treatment of cerebral spasticity preoperative electric stimulation of posterior roots and its use in the choice of the roots to be sectioned. Neurochirurgie 22, 23 – 34. Fyffe, R.E., Light, A.R., 1984. The ultrastructure of group Ia afferent fiber synapses in the lumbosacral spinal cord of the cat. Brain Res. 300, 201 – 209. Garthwaite, J., 1991. Glutamate, nitric oxide and cell – cell signaling in the nervous system. Trends Neurosci. 14, 60 – 67. Gerebtzoff, M.A., Maeda, T., 1968. Caracte`res et localisation histochimique d’un isoenzyme fluorore´sistant de la phosphatase acide dans la moelle e´pinie`re du rat. C. R. Soc. Biol. (Paris) 162, 2032 – 2035. Gonza´lez-Herna´ndez, T., Rustioni, A., 1999a. Expression of three forms of nitric oxide synthase in peripheral nerve regeneration. J. Neurosci. Res. 55, 198 – 207. Gonza´lez-Herna´ndez, T., Rustioni, A., 1999b. Nitric oxide synthase and growth-associated protein are coexpressed in primary sensory neurons after peripheral injury. J. Comp. Neurol. 404, 64 – 74.

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178 Grozdanovic, Z., Baumgarten, H.G., 1999. Nitric oxide synthase in skeletal muscle fibers: a signaling component of the dystrophin – glycoprotein complex. Histol. Histopathol. 14, 243 – 256. Hal*a´k, L., Pecha´nˇova´, O., Zˇigova´, Z., Klemova´, L., Novacky´, M., Berna´tova´, I., 2000. Inhibition of NO synthase activity in nervous tissue leads to decreased motor activity in the rat. Physiol. Res. 49, 143 – 149. Ho¨kfelt, T., Kellerth, J.O., Nilssen, G., Pernow, B., 1975. Substance P: localisation in the central nervous system and in some primary sensory neurons. Science 19, 889 – 890. Hongo, T., Kudo, N., Sasaki, S., Yamashita, M., Yoshida, K., Ishizuka, N., Mannen, H., 1987. Trajectory of group Ia and Ib fibers from the hindlimb muscles at the L3 and L4 segments of the spinal cord of the cat. J. Comp. Neurol. 262, 159 – 194. Iles, J.F., 1986. Reciprocal inhibition during agonist and antagonist contraction. Exp. Brain Res. 62, 212 – 214. Ishizuka, N., Mannen, H., Hongo, T., Sasaki, S., 1979. Trajectory of group Ia afferent fibers stained with horseradish peroxidase in the lumbosacral spinal cord of the cat: three dimensional reconstructions from serial sections. J. Comp. Neurol. 186, 189 – 212. Kausz, M., Re´thelyi, M., 1985. Lamellar arrangement of neuronal somata in the dorsal root ganglion of the cat. Somatosens. Res. 2, 193 – 204. Kawabata, A., Fukuzomi, Y., Fukushima, Y., Takagi, H., 1992. Antinociceptive effect of l-arginine on the carrageenin-induced hyperalgesia of rat: possible involvement of central opioidergic systems. Eur. J. Pharmacol. 218, 153 – 158. Kawabata, A., Manabe, S., Manabe, Y., Takagi, H., 1994. Effect of topical administration of l-arginine on formalin-induced nociception in the mouse: a dual role of peripherally formed NO in pain modulation. Br. J. Pharmacol. 112, 547 – 550. Kitto, K.F., Haley, J.E., Wilcox, G.L., 1992. Involvement of nitric oxide in spinally mediated hyperalgesia in the mouse. Neurosci. Lett. 148, 1 – 5. Knyiha´r, E., 1971. Fluoride-resistant acid phosphatase system of nociceptive dorsal root afferents. Experientia 27, 1205 – 1207. Kuno, M., Munoz-Martinez, E.J., Randic, M., 1973. Synaptic action on Clarke’s column neurones in relation to afferent terminal size. J. Physiol. 228, 343 – 360. Laitinen, L.V., Fugl-Meyer, A.R., 1982. Assessment of functional effect of epidural electrostimulation and selective posterior rhizotomy in spasticity. Appl. Neurophysiol. 45, 33l – 334. Laitinen, L.V., Nilsson, S., Fugl-Meyer, A.R., 1983. Selective posterior rhizotomy for treatment of spasticity. J. Neurosurg. 58, 895 – 899. Lazorthes, Y., Sol, J.-C., Sallerin, B., Verdie´, J.-C., 2002. The surgical management of spasticity. Eur. J. Neurol. 9, 35 – 41. Lieberman, A.R., 1976. Sensory ganglia. In: London, D.N. (Ed.), The Peripheral Nerve. Chapman and Hall Pub’s, London, pp. 188 – 278. Luka´*ova´, N., Cˇ´ı”kova´, D., Marxala, M., Jal*, P., Marxala, J., 1999. Segmental and laminar distributions of nicotinamide adenine dinucleotide phosphate-diaphorase-expressing and neuronal nitric oxide synthase-immunoreactive neurons versus radioassay detection of catalytic nitric oxide synthase activity in the rabbit spinal cord. Neuroscience 94, 229 – 237. Luka´*ova´, N., Cˇ´ı”kova´, D., Kri”anova´, O., Pavel, J., Marxala, M., Marxala, J., 2003. Peripheral axotomy affects nicotinamide adenine dinucleotide phosphate diaphorase and nitric oxide synthases in the spinal cord of the rabbit. J. Neurosci. Res. 71, 300 – 313. Luka´*ova´, N., Kolesa´r, D., Marxala, M., Marxala, J., in press. Immunohistochemical, histochemical and radioassay analysis of nitric oxide synthase immunoreactivity in the lumbar and sacral dorsal root ganglia of the dog. Cell. Mol. Neurobiol. Malmberg, A.B., Yaksh, T.L., 1993. Spinal nitric oxide synthesis inhibition blocks NMDA induced thermal hyperalgesia and produces antinociception in the formalin test in rats. Pain 54, 291 – 300. Manjarrez, E., Teresa, R., Rojas-Piloni, G., Mendez, I., Velez, D., Flores, A., 2001. NO donor SIN-1 potentiates monosynaptic reflexes in the cat spinal cord. NeuroReport 12, 2667 – 2671.

177

Marxala, J., Sˇulla, I., Jalcˇ, P., Orenda´*ova´, J., 1995. Multiple protracted cauda equina constrictions cause deep derangement in the lumbosacral spinal cord circuitry in the dog. Neurosci. Lett. 193, 97 – 100. Marxala, J., Kluchova´, D., Marxala, M., 1997. Spinal cord gray matter layers rich in NADPH diaphorase-positive neurons are refractory to ischemia-reperfusion-induced injury: a histochemical and silver impregnation study in rabbit. Exp. Neurol. 145, 165 – 179. Marxala, J., Vanicky´, I., Marxala, M., Jal*, P., Orenda´*ova´, J., Taira, Y., 1998. Reduced nicotinamide adenine dinucleotide phosphate diaphorase in the spinal cord of dogs. Neuroscience 85, 847 – 862. Marxala, J., Marxala, M., Vanicky´, I., Taira, Y., 1999. Localization of NADPHd-exhibiting neurons in the spinal cord of the rabbit. J. Comp. Neurol. 406, 263 – 284. Marxala, J., Luka´*ova´, N., Cˇ´ı”kova´, D., Kafka, J., Katsube, N., Kucha´rova´, K., Marxala, M., 2002. The case for the bulbospinal respiratory nitric oxide synthase-immunoreactive pathway in the dog. Exp. Neurol. 177, 115 – 132. Marxala, J., Marxala, M., Luka´*ova´, N., Ishikawa, T., Cˇ´ı”kova´, D., 2003. Localization and distribution patterns of nicotinamide adenine dinucleotide phosphate diaphorase exhibiting axons in the white matter of the spinal cord of the rabbit. Cell. Mol. Neurobiol. 23, 57 – 92. Marxala, J., Luka´*ova´, N., Cˇ´ı”kova´, D., Luka´*, I., Kucha´rova´, K., Marxala, M., 2004. Premotor nitric oxide synthase immunoreactive pathway connecting lumbar segments with the ventral motor nucleus of the cervical enlargement in the dog. J. Chem. Neuroanat. 27, 43 – 54. Maxwell, D.J., Christie, W.M., Ottersen, O.P., Storm-Mathisen, J., 1990. Terminals of group Ia primary afferent fibers in Clarke’s column are enriched with l-glutamate-like immunoreactivity. Brain Res. 510, 346 – 350. Mazzocchio, R., Rossi, A., 1997. Involvement of spinal recurrent inhibition in spasticity. Further insight into the regulation of Renshaw cell activity. Brain 120, 991 – 1003. Meller, S.T., Gebhart, G.F., 1994. The role of nitric oxide in hyperalgesia. In: Urban, L. (Ed.), The Cellular Mechanisms of Sensory Processing NATO ASI Series, vol. H79. Springer, Heidelberg, pp. 401 – 419. Meller, S.T., Dykstra, C., Gebhart, G.F., 1992a. Production of endogenous nitric oxide and activation of soluble guanylate cyclase are required for N-methyl-d-aspartate-produced facilitation of the nociceptive tail-flick reflex. Eur. J. Pharmacol. 214, 93 – 96. Meller, S.T., Pechman, P.S., Gebhart, G.F., Maves, T.J., 1992b. Nitric oxide mediates the thermal hyperalgesia produced in a model of neuropathic pain in the rat. Neuroscience 50, 7 – 10. Meller, S.T., Dykstra, C., Gebhart, G.F., 1993. Characterization of the spinal mechanisms of thermal and mechanical hyperalgesia following intraplantar zymosan. Abstr. -Soc. Neurosci. 19, 967. Meller, S.T., Cummings, C.J., Traub, R.J., Gebhart, G.F., 1994. The role of nitric oxide in the development and maintenance of hyperalgesia and edema produced by intraplantar carrageenan in the rat. Neuroscience 60, 367 – 374. Molander, C., Ygge, J., Dalsgaard, C.J., 1987. Substance P-somatostatinand calcitonin gene-related peptide-like immunoreactivity and fluoride resistant acid phosphatase-activity in relation to retrogradely labeled cutaneous, muscular and visceral primary sensory neurons in the rat. Neurosci. Lett. 74, 37 – 42. Moncada, S., Higgs, A., 1993. The l-arginine-nitric oxide pathway. N. Engl. J. Med. 329, 2002 – 2012. Morita, H., Crone, C., Christenhuis, D., Petersen, N.T., Nielsen, J.B., 2001. Modulation of presynaptic inhibition and disynaptic reciprocal Ia inhibition during voluntary movement in spasticity. Brain 124, 826 – 837. Nagy, J.I., Hunt, S.P., 1982. Fluoride-resistant acid phosphatase-containing neurons in dorsal root ganglia are separate from those containing substance P or somatostatin. Neuroscience 7, 89 – 97. Nicol, M.J., Walmsley, B., 1991. A serial section electron microscope study of an identified Ia afferent collateral in the cat spinal cord. J. Comp. Neurol. 314, 257 – 277. Nielsen, J., Petersen, N., Crone, C., 1995. Changes in transmission across synapses of Ia afferents in spastic patients. Brain 118, 995 – 1004.

178

J. Marsˇala et al. / Experimental Neurology 195 (2005) 161 – 178

Orenda´cˇova´, J., Marxala, M., Sˇulla, I., Kafka, J., Jal*, P., Cˇ´ı”kova´, D., Taira, Y., Marxala, J., 2000. Incipient cauda equina syndrome as a model of somatovisceral pain in dogs. Spinal cord structures involved as revealed by the expression of c-fos and NADPH diaphorase activity. Neuroscience 95, 547 – 557. Orenda´*ova´, J., Cˇ´ı”kova´, D., Kafka, J., Luka´*ova´, N., Marxala, M., Sˇulla, I., Marxala, J., Katsube, N., 2001. Cauda equina syndrome. Prog. Neurobiol. 64, 613 – 637. Pierce, J.P., Mendell, L.M., 1993. Quantitative ultrastructure of Ia boutons in the ventral horn: scaling and positional relationships. J. Neurosci. 13, 4748 – 4763. Ramon y Cajal, S., 1952. Histologie du Syste`me Nerveux de l’Homme et des Verte´bre´s. Instituto Ramon y Cajal, Madrid. Re´thelyi, M., 1970. Ultrastructural synaptology of Clarke’s column. Exp. Brain Res. 11, 159 – 174. Rosenfeld, M.G., Mermod, J.-J., Amara, S.G., Swanson, L.W., Sawchenko, P.E., Rivier, J., Vele, W.W., Evans, R.M., 1983. Production of novel neuropeptide encoded by the calcitonin gene via tissue-specific RNA processing. Nature 304, 129 – 135. Saito, K., 1974. The synaptology and cytology of the Clarke cell in nucleus dorsalis of the cat: an electron microscopic study. J. Neurocytol. 3, 179 – 197. Saito, S., Kidd, G.J., Trapp, B.D., Dawson, T.M., Bredt, D.S., Wilson, D.A., Traystman, R.J., Snyder, S.H., Hanley, D.F., 1994. Rat spinal cord neurons contain nitric oxide synthase. Neuroscience 59, 447 – 456. Scheibel, M.E., Scheibel, A.B, 1969. Terminal patterns in cat spinal cord: III. Primary afferent collaterals. Brain Res. 13, 417 – 443. Scherer-Singler, U., Vincent, S., Kimura, H., McGeer, E.G., 1983. Demonstration of a unique population of neurons with NADPHdiaphorase histochemistry. J. Neurosci. Methods 9, 229 – 234. Sindou, M., Quoex, C., Baleydier, C., 1974. Fiber organization at the posterior spinal cord-rootlet junction in man. J. Comp. Neurol. 153, 15 – 26. Snyder, R., 1977. The organization of the dorsal root entry zone in cats and monkeys. J. Comp. Neurol. 174, 47 – 70. Sommer, E.W., Kazimierczak, J., Droz, B., 1985. Neuronal subpopulations in the dorsal root ganglion of the mouse as characterized by

combination of ultrastructural and cytochemical features. Brain Res. 346, 310 – 326. Sousa, A.M., Predo, W.A., 2001. The dual effect of a nitric oxide donor in nociception. Brain Res. 897, 9 – 19. Szarijanni, N., Re´thelyi, M., 1979. Differential distribution of small and large neurons in the sacrococcygeal dorsal root ganglia of the cat. Acta Morphol. Acad. Sci. Hung. 27, 25 – 35. Szenta´gothai, J., 1958. The anatomical basis of synaptic transmission of excitation and inhibition in motoneurons. Acta Morphol. Acad. Sci. Hung. 8, 287 – 309. Szenta´gothai, J., Albert, A., 1955. The synaptology of Clarke’s column. Acta Morphol. Acad. Sci. Hung. 5, 43 – 51. Tasci, N., Ankarali, S., Demir, S., 2003. Further evidence for enhancing effects of NO on monosynaptic and polysynaptic spinal reflexes in cats. Brain Res. 980, 109 – 116. Terenghi, G., Riveros-Moreno, V., Hudson, L.D., Ibrahim, N.B., Polak, J.M., 1993. Immunohistochemistry of nitric oxide synthase demonstrates immunoreactive neurons in spinal cord and dorsal root ganglia of man and rat. J. Neurol. Sci. 118, 34 – 37. Thompson, S.W., Babbedge, R., Levers, T., Dray, A., Urban, L., 1995. No evidence for contribution of nitric oxide to spinal reflex activity in the rat spinal cord in vitro. Neurosci. Lett. 188, 121 – 124. Tracey, D.J., Walmsley, B., 1984. Synaptic input from identified muscle afferents to neurones of the dorsal spinocerebellar tract in the cat. J. Physiol. (London) 350, 599 – 614. Valtschanoff, J.G., Weinberg, R.J., Rustioni, A., 1992. NADPH diaphorase in the spinal cord of rats. J. Comp. Neurol. 321, 209 – 222. Walmsley, B., Wienawa-Narkiewicz, E., Nicol, M.J., 1985. The utrastructural basis for synaptic transmission between primary muscle afferents and neurons in Clarke’s column of the cat. J. Neurosci. 8, 2095 – 2106. Wang, M.X., Murrell, D.F., Szabo, C., Warren, R.F., Sarris, M., Murrell, G.A.C., 2001. Nitric oxide in skeletal muscle: inhibition of nitric oxide synthase inhibits walking speed in rats. Nitric Oxide 5, 219 – 232. Wilson, L.B., Engbretson, J., 1998. Effect of NO in the dorsal horn on the pressor reflex evoked by muscle mechanoreceptor activation. Abstr. Soc. Neurosci. 24, 622.