- Email: [email protected]
Immunohistochemical localization of calretinin in the dorsal root ganglion and spinal cord of the rat
I~rcrrnRPwarcll Bll//r~lrrl.Vol. 3 I. pp. 13-2, Printed in the LISA. All rights reserved.
03hl-Y730/93 $6.00 + .oo Copyright c I993 Pergamon Press Ltd.
1993
Immunohistochemical Localization of Calretinin in the Dorsal Root Ganglion and Spinal Cord of the Rat K.
*Neurobiology Science,
REN.*’
M. A. RUDA*
AND
D. M. JACOBOWITZf
and Anesthesiology Branch, National Institute qf Dental Research and ilahoratoq~ ofClinical National Institute qj”A4ental Health, National Institutes of’ Health. Bethesda. MD 20892
Received
17 August
1992; Accepted
26 August
1992
REN. K.. M. A. RUDA AND D. M. JACOBOWITZ. Irnmllnohisto~hemic,ul loulimion of‘cdretinin uncl spind cord q/‘thc rut. BRAIN RES BULL 31( l/2) 13-22. 1993.-Calretinin (CR). a recently
in thr dord
root ~unglron
identified calcium-binding protein. is present in nervous tissue, including sensory pathways. where it may play an important role in regulation of cellular activity. Using immunocytochemistry, we examined the cellular localization of CR in dorsal root ganglia (DRG) and spinal cord of normal rats and after multiple unilateral dorsal root ganglionectomies. In DRG. CR-immunoreactive cell bodies and axons were a small subpopulation (10%) ofmediumto large-sized neurons. In the spinal cord. CR-like immunoreactivity (LI) in neurons and fibers was found in all laminae except motoneurons. Dense fiber networks were also found in Clarke’s column. The densest staining of both cell bodies and fibers was in the superficial laminae. especially lamina II. and in the lateral spinal and lateral cervical nuclei. CR-immunoreactive fibers were also observed in the fasciculi cuneatus and gracilis. Fasciculus gracilis exhibited the greatest number of labeled axons at the lumbosacral levels, but few labeled axons were found at the rostra1 thoracic and cervical levels. In contrast. the corticospinal tract at the base of the dorsal column was devoid of CR-immunoreactive fibers. Unilateral multiple lumbar ganglionectomies resulted in a loss of CR-L1 in the dorsal columns ipsilateral to the surgery. In the spmal gray matter ipsilateral to the ganglionectomies. CR-L1 was reduced in Clarke’s column and slightly enhanced in the medial third of lamina II. Our observations demonstrate a unique distribution pattern of CR-L1 compared to other calcium-binding proteins in the spinal cord. and suggest a role for CR in nociceptive and proprioceptive pathways. Spinal cord Calretinin Calcium-binding proteins
Dorsal root ganglion Rat
Dorsal column
(CR) is a newly discovered member of the family of calcium-binding proteins (I 9.24). It is closely related to another calcium-binding protein, calbindin D-28K (3J.14). The primary translation product of CR mRNA is 29 kDa (19). CR and calbindin D-28K have a 50-60s cDNA and amino acid sequence homology ( 14.15.19.24). RNA blots have detected CR mRNA in the brain but not in peripheral tissues ( 15,I9). Thus, CR appears to he primarily a neuronal protein. CR contains six potential calcium binding domains, five of which are thought to be active in binding calcium (14). Although it seems obvious that CR may modulate calcium concentrations in cells, its functional significance has not been determined. It is thought that CR, as well as some other calcium-binding proteins, perform unknown functions other than just buffering cellular calcium concentrations (2 I). This view is supported by the specific and differential distribution patterns of calcium-binding proteins in the nervous system (1.8.17-19,26,27). It has been
should
be addressed
Immunocytochemistry
recently shown that CR is involved in the regulation of protein phosphorylation in the brain (25). Although the distribution of CR immunoreactivity in the rat and chick brain has been studied in great detail. a detailed account of the distribution in the spinal cord and dorsal root ganglia has not been reported (1,8,18,20). Using a specific antiserum raised against guinea pig CR. originally named protein 10 (24). a detailed description of CR-like immunoreactivity (LI) in the cervical through lumbosacral spinal cord is presented in this report. We also describe the localization of CR-L1 in a subpopulation of dorsal root ganglion (DRG) neurons. The effects of unilateral multiple ganglionectomies on CR-L1 in the spinal cord were also studied.
CALRETININ
’ Requests for reprints 20892.
Ganglionectomy
METHOD
Ten male Sprague-Dawley rats (200-300 the present study. Five rats received unilateral
to K. Ren. Ph.D. NAB/NlDR/NIH
13
Building
30. Room
B-20. 9000 Rockville
g) were used in multiple dorsal
Pike. Bethesda,
MD
14
root gan~ionectomies under sterile and anesthetic conditions and were allowed to survive one (n = 2) to two (n = 3) weeks after the surgery. All rats were deeply anesthetized with sodium pentobarbital and perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. The extent of the ganglionectomies was verified after the perfusion. The cervical, thoracic, lumbar and sacral spinal cord and DRG were removed, immersed in the same fixative overnight at 4”C, and transferred to 30% sucrose (w/v) in phosphate buffer for several days for cryoprotection. Thirty micron-thick sections were cut with a cryostat at -20°C in either the transverse or sagittal plane. The sections were collected in phosphate buffered saline (PBS) and processed for immunocytochemistry. The primary antiserum against guinea pig CR was raised in rabbits as described (24). The specificity of the antiserum has been examined previously (1,8). Incubation of CR antiserum with purified CR eliminated immunostaining of the brain sections (1,s). Free-floating tissue sections were rinsed in PBS with 0.75% Triton X-100 for I h, 5%) normal goat serum (NGS) in PBS for 30 mitt, then incubated with anticalretinin antiserum ( I :7500- 10,000 in PBS) with I-3% NGS for 48-72 h with gentle agitation. After two washes in PBS (10 min each), and a 20-min rinse in PBS with 5% NGS, the sections were incubated with biotinylated goat antirabbit IgG (Vector, I:200 in PBS) for 30-60 min. Following rinses in PBS (10 min X 2) and PBS with 5% NGS (20 min), the sections were then incubated with avidin and biotinylated horseradish peroxidase complex (Vector, 1:50 in PBS) for 30-60 min. After washes in PBS or Tris HCl-buffered saline (TBS), the tissue sections were reacted with 0.05% diaminobenzadine dihydrochloride (DAB, Sigma) in 0.1 M phosphate buffer containing 0.003% hydrogen peroxide for 6 min. In some experiments, the sections were preincubated in 0.2% nickel ammonium sulfate in 0.1 M phosphate buffer for 10 min before incubation with DAB. Both procedures produced the same immunostaining pattern in the tissue although the nickel ammonium sulfate incubation resulted in increased contrast. Some sections were later counterstained with cresyl violet to allow visualization of the general histology of the tissue. Control sections were processed with the same method except that the primary antiserum was omitted. The sections were finally rinsed in PBS, mounted on gelatin-coated slides, dehydrated in alcohols, cleared in xylenes, and coverslipped with Eukitt (Kindler, Germany). Immunostaining intensity and packing density of CR-immunoreactive (IR) cell bodies and fibers were subjectively assessed as described (1,8). The morphology of some spinal neurons was traced with a camera lucida. CR-L1 also was analyzed with a computer-assisted image analysis system (Image 1.41, National Institutes of Health). The size of labeled neurons was determined by tracing the outlines of cell bodies and calculating their cross-sectional areas (pm’). An effort was made to measure neurons with a visibfe nucleus and nucleolus. When nucleolus was indistinguishable, especially for small neurons, neurons with clear nuclei were measured. Spinal neurons (n = 1736) were measured in randomly selected sections from five animals and arbitrarily divided into three groups according to their sizes; that is, small, < 150 pm2; medium, 150-300 pm*; and large, > 300 Frn’ ( 1). DRG neurons were usually larger and thus grouped differently: small, < 1000 Frn’; medium, 1000-2000 pm*: and large, > 2000 pm* (10). Rat spinal cord structures were identified and named according to the stereotaxic atlases of Paxinos and Watson ( 16) and Molander et al. f 12.13).
RI-N. RL;DA /ihI)
JAC‘OROWI I /
RtSl.II.IS
The CR-L1 was observed at all segmental levels studied. Some segmental differences, especially in specialized nuclear groups. were found. For instance, the CR-IR cell bodies at the lumbosacral spinal cord were more abundant than those at the ccrvicothoracic spinal cord. The densest immunoreactive staining was found in the superficial dorsal horn, especially lamina II. and the lateral spinal and lateral cervical nuclei. Examples of the CR-L1 in the representative segments of the spinal cord are illustrated in Fig. I. An analysis of the CR-IR cell bodies and fibers is summarized in Table 1. A few examples of CR-IR neurons are shown in Fig. 2. Many nuclei were stained denser than the cytoplasm (Fig. 2C, D). The nucleolus was not labeled (Fig. 2A, E). Glia cells were not stained. Dorsul root ganglion. Approximately 10% of total number of neurons in the DRG were labeled with CR antiserum. No segmental difference was found at cervical through sacral levels. Variable staining density, not correlated with cell size, was found (Fig. 3). Cross-sectional area of the labeled neurons ranged from 300 to over 3000 pm*, although much of the cells were medium to large sized (90%). Labeled axons, especially with large diameters (>2 gm). were found in the ganglia and dorsal roots. Axons originating from cell bodies could be easily seen (Fig. 3B). The CR-IR axons could be traced into the dorsal root entry zone of the spinal cord (Fig. 4). ~F?z~nu I. The largest lamina I cells, very likely co~esponding to Waldeyer cells or marginal ceils [see (23), for a review], were labeled. In transverse sections, Waldeyer cells were seen under Lissauer’s tract or in the middle third of lamina I. They typically have a mediolaterally oriented fusiform cell body with dendrites extending along the border of lamina I (Fig. 2B). The crosssectional areas of the Waldeyer cell bodies were 100-300 mm*,two to three times larger than lamina II CR-IR neurons. In addition to Waldeyer cells. smaller round cell bodies were also CR-IR. A CR-IR plexus of nerve fibers was found in lamina I. However, it did not appear to be quite as dense as that in lamina II. L~r7ziF~uII. A densely stained band, corresponding to lamina II in the superficial dorsal horn, was found in all segments of the cord examined (Fig. 1). The immunor~ctive materials consisted of high density neuropil and densely packed cell bodies (Figs. 2B and 4). The fiber plexus staining intensity was approximately the same for outer and inner portions of lamina II through most spinal segments, However, CR-IR ceil bodies were densely packed along the border of laminae II and III in the rostra1 cervical cord. At different segment levels, CR exhibited variable density between the lateral and medial aspects of lamina II. At cervical levels, lateral lamina II was typically more densely stained than medial lamina II. At lumbar levels, the pattern was similar. although CR-L1 was generally denser than at cervical levels. In Contras& at both thoracic and sacral levels, CR-L1 was comparatively more dense from the lateral to the medial part of Iamina II (Fig. I). Cross-sectional areas of most Iamina II CR-IR neurons were smaller than 100 ym2. In the transverse plane, the ceil bodies appeared as round or oval in shape and dendritic orientations were not clear, In sagittal sections, however, many lamina II CR-IR dendrites were observed extending rostrocaudahy (Figs. 1E and SA). One type of lamina II CR-IR neuron, mainly found in outer lamina II, had ventrally directed dendrites (Fig. SB), resembling those of the limiting cell or stalked cell described previously [(2,7), also see (4), for a review]. The other type of CR-IR neuron located in both the outer and inner portion of lamina II resembled the central cell or islet cell [(2,7), also see (4). for a review], whose
CAL.RETININ
IN THE RAT SPINAL CORD
FIG. 1. The dist~bution of cairetinin-like iummunoreacfivity in the cervical C5 (A). lumbar L5 (B), thoracic T3 (C). and sacral S2 (If) spinal cord. The densest staining was found in the superficial dorsal horn and the lateral spinal nucleus. Note that the immunoreactive bands in the superficiai dorsal horn are less dense medially at the lumbar and cervical levels. Also note the differential staining pattern in the dorsal columns at cervical, thoracic, lumbar, and sacral levels. Laminae 111and IV. pyramidal tract (PT), and ventral root (e.g., arrow heads in A) generally lack CR-LI. A sagittai section from the lumbar enlargement, cut through a sagittal plane indicated by an arrow in B, is shown in E. CC, Clarke’s column; CU, fasciculus cuneatus; DC, dorsal column; DGC, dorsal gray commissure; GR, fasciculus gracilis; IML, intermediolateral nucieus; LSN, lateral spinal nucleus. Roman numerals I-X indicate spinal cord laminae. (A-D) Nickel enhanced DAB reaction; (E) DAB reaction. Scale bar (same for A-E) = 200 pm, X37.5.
dendrites mainly travel rostrocaudally. Occasionally, the recurrent process on these neurons was Seen (Fig. 5A). Laminuc III und IV. This area of the spinal cord gray matter consistently exhibited the least CR-LI. Only few scattered cell bodies and axons were labeled. The neuropil in this region was generally devoid of CR-LI. Laminae L’ and W. Dense immunoreactive neuropil and numerous labeled cells were found in these laminae. Most immunoreactive neurons were medium to large multipolar neurons
(Fig. I B. E). The immunoreactive
reaction product was evenly distributed in the cytoplasm. The dendrites and axons could be traced up to 200-400 pm away from the cell bodies. An example of a lamina VI multipolar neuron is shown in Fig. 2A. Pyramidal and smail bipolar neurons also were seen. Laminae V-VI neurons with CR-LI were more abundant at the lumbar level than at cervical, thoracic, and sacral levels. At the lumbar level, immunoreactive cell bodies and neuropil were mainly distributed in the middle third of the iaminae (Fig. 1B).
IABLE ~MMUNOHIST~HE~ICAL
LOCALIZATION ---~-
1
OF CALRETININ
IN THE RAT SPINAL
(‘ell Bodies Spinal Cord Region
Lamina I Lamina II Laminae III/IV Lamina V/VI Cervicothoracic Lumbo~~ral Lamina VII Cervicothoraeic Lumbosacral Lamina VIII Lamina IX Lamina X Clarke’s column Lissaeur’s tract Lateral spinal nucleus Gracile fasciculus LumbosacraI Cervicothoracic Cuneate faseiculus Pyramidal tract
Shape*
Size?
Staimng Intensityt
CORD
__...Packing Densityg
I-ibcn __
Packing Density*
0, f
S. M
t, ++
0. f 0. f, m
s s
++ +. ++
f. P, m f. p. m
S. M, l..
+. ++
+
ti
S. M, L
f. if
C-F, +++
++, +++
f, m f, m f. m
M. L M. L M. 1.
+, ++ +. ++ +. ++
+
+. ++
t, t+ t
0. f
S. M
+, ++
-t -
0, f
S
f, ++
+-, +
++. t ++ i t, t-i + +t. i-i+ i tt -I-
if,
ir -t-t+
-
i -! r
+ f. i +-t. ft fS-. i-i+,
ft
* o, oval or round; f, fusiform; p, pyramidal; m, muIti~Iar. i S, < I50 pm’: M, 150-300 pm’; L, >3OO pm’; (rm2, cell crosssectional area). $ +t-, intense; +, light. 9:+++, high density; ++, moderate density; t, low density; +, rarely found; -, not found.
# +++, high-density neuropil; ++, moderate-density neuropil; +, low-density neuropil; -. not found; ft. fiber tract.
Laminae Hi and VIII. Many medium- to large-size neurons with a variety of shapes were stained. Multipolar neurons were the most abundant in this intermediate region of the spinal cord gray matter. The cross-sectional area ranged from 100 to i000 pm’. Their dendrites were inte~in~~, and most of them apneared to be oriented do~oventrally when viewed in the sagittal plane (Fig. I E). Lamina IX. Although CR-L1 was present in fiber networks, the large motoneurons in lamina IX were not labeled by CR antiserum. This was clearly shown in sections counterstained with cresyl violet (data not shown). Consistently, the ventral roots originating from motoneuron pools were not labeled (Fig. IA). Lamina X. CR-IR cell bodies were frequently found dorsal to the central canal. Occasionally, CR-IR cell bodies were found lateral to the central canal, but essentially no labeled neurons were found ventral to the central canal. Most labeled neurons were small. A few medium flattened cells with cell bodies oriented m~olaterally were seen in this region. Densely packed immunoreactive cell bodies were clustered in the dorsal gray commissure (DGC) of the sacral spinal cord. Their processes were distributed horizontally, reaching the intermediolateral nucleus. Ceil bodies in the DGC also sent processes running dorsally towards the edge ofthe gray matter. A typical section illustrating CR-L1 in the DGC is shown in Fig. ID. Other special spinal nuclei. At the upper cervical spinal cord (Cl-3) immunoreactive fiber networks were aggregated in the territories of the internal basilar nucleus and central cervical nucleus. Clarke’s column was clearly distinguishable at the thoracic level by a densely labeled neuropil (Fig. 1C). Very few CR-
IR cell bodies were found in the central cervical nucleus, Clarke’s column, intermediate medial nucleus, and intermediate lateral nucleus at the related segments. White matter. The white matter surrounding the spinal cord contained numerous CR-IR axons. The dorsal lateraf fascicufus was densely stained. A few labeled cell bodies were occasionally seen in the Iateral spinal nucleus and lateral cervical nucleus. In some sections, elongated neurons with dendritic processes extending from the lateral spinal nucleus to the dorsal horn could be seen. In Lissauer’s tract, few fine fibers were stained (Fig. 4). Many fibers in the dorsal column fasciculi cuneatus and gracilis were immunoreactive to the calretinin antiserum. Interestingly, although clear immunostaining was found in the fasciculus gracilis at the Iumbosacral levels, very few immunoreactive fibers were observed in the fasciculus gracihs at the rostra1 thoracic and cervical levels (Fig. I). The fasciculus cuneatus was usually full of immunoreactive fibers, but at the CI level, lateral part of the fasciculus cuneatus dorsal to the medial portion ofthe dorsal horn also was not stained. Most parts of the ventrat and lateral white matter were stained by CR antiserum. The pyramidal tract in the ventromedial portion of the dorsal coiumn and fibers in the dorsal midline area (tail zone) were not CR-IR. Efects of UnilateralMultiple Dorsal Root Ganglionectomies Five rats received unilateral multiple ganglionectomies primarily at the lumbar levels. The locations ofthe ganglionectomies were L2-5, L I-5, L3-S I, and L2-6. The other one rat had L I 2 ganghonectomies and L3 and L6 dorsal root axotomies. The ganghonectomized rats were perfused 1 (n = 2) or 2 (n = 3) weeks after the surgery.
CALRETININ
IN THE
RAT SPINAL
17
CORD
FIG. ?. Photomicrographs illustrating calretinin-immunoreactive cell bodies in the spinal cord. (A) A multipolar neuron (arrow) in lamina Vi. Note uniform immunostaining in the cytoplasm, nucleus, and processes. The nucleolus, the clear dot in the center of the cell. was not stained. Arrowheads indicate an axon-like process. (B) A Waldeyer type lamina I neuron (arrow) in the sacral spinal cord with mediolateraily extended dendrites (arrowheads). Another large round lamina I neuron (asterisk) is also present in this section. Numerous CR-IR cell bodies are also found in lamina II. (C) A multipolar CR-IR neuron (arrow) in lamina VII. Note that the nucleus is stained more densely than the cytoplasm. (D) A pyramidal neuron (arrow) in lamina V. The nucleus is densely stained. (E) Arrow indicates a bipolar CR-IR neuron in lamina VIII. Nickel enhanced DAB reaction. Scale bar (same for 4-E) : 25 pm. X316
There was a loss of CR-L1 in the dorsal column fascicuius gracilis at related segments ipsilateral to the ganglionectomy or axotomy. L6 dorsal root axotomy with L4, 5 DRG intact resulted in a loss of CR-LI in the dorsoiaterai corner of the ipsilateral dorsal column at the related segment. L2-5 ganghonectomies essentially led to a total loss of immunostaining in the ipsiiaterai dorsal column (Fig. 6). The changes in spinal gray matter were less apparent after gangiionectomy. CR-L1 was reduced slightly in the nucleus dorsalis (Clark’s column) rostrai to the surgery levels. The CR-L1 was slightly enhanced in the medial third of lamina 11 at the surgical level ipsiiaterai to the gangiionectomy. In four animals with at least L4, 5 gangiionectomies, the staining intensity and packing density of CR-IR cell bodies was slightly increased especially in the medial end of iamina II (Fig. 6).
DISCUSSION A distinct and reproducible CR-L1 was observed in the cervical through sacral DRG and spinal cord of the rat in the present study. The CR-L1 was extensively distributed in spinal cord structures including iaminae I, 11, V-VIII, and X, the lateral spinal and lateral cervical nuclei, and the dorsal column fascicuii cuneatus and gracilis. The CR-L1 was found in the neuronal cytoplasm. nucleus. dendrites, and axons. The immunocytochemical localization of CR in the spinal cord has been briefly described by Rogers (20) and Resibois and Rogers (18). The present results are generally consistent with their results. It was further found that the dorsal root ganglion~tomy primarily resuited in a loss of CR-L1 in the dorsal column. The CR-L1 in spinal gray matter was largely preserved after the gangiionectomy. Obviously. although CR-IR primary afferents project through
FIG. 3. Dorsal root ganglion neurons stained with calretinin antiserum. (A) The neurons exhibit both dense (arrows) and moderate (arrowheads) cytoplasmic immunoreactivity. Most CR-IR cells are medium to large sized. (B) Arrowheads point to a CR-IR axon as it exits from the neuronal cell body. DAB reaction. Scale bars: A. 100 pm, X100: B. 25 pm. x380.
the spinal cord, few CR-IR primary afferents terminating in the spinal gray matter, and spinal neurons exhibit intrinsic CR-LI. The differential distribution patterns of different calciumbinding proteins are most impressive. It is apparent from this study that CR is differentially distributed in the spinal cord. when compared with the locakation of other calcium-binding proteins. CR was most abundandy distributed in iamina II and the lateral spinal and lateral cervical nuclei of the spinal cord. The CR-L1 in laminae III, IV, and IX was either very light or absent. In contrast, parvalbumin is primarily localized in inner lamina II and outer lamina III of the spinal cord (26.27). A close
relative of CR, calbindin D-28K, is most abundant in laminae 1and II. Its distribution pattern is homogeneous through different spinal segments as compared to CR (27). In contrast to CR, caimodulin-LI was not detectable in lamina I1 of the dorsal horn, but did exist in the motoneuron pools ofthe ventral horn (lamina IX) (unpublished observations). The specific and differential localization of calcjum-binding proteins has also been noted in the brain [for examples, ( 1,8,1Xl]. The functional significance of differential or complementary localization of different calcium binding proteins in the nervous system is not understood. It is reasonable to suggest that each
FIG. 4. Calretinin-like immunoreactivity in the dorsal root entry zone in a transverse section from CS spinal cord. Thick lab&d fibers can be traced into the dorsal column (arrowheads) from the dorsal root. Few fine labeled fibers are seen in Lissauer’s tract (LISS). DR, dorsal root: CU. fasciculus cuneatus. Nickel enhanced DAB reaction. Scale bar = 50 pm* X200.
(‘ALRETININ
I9
IN THE RAT SPINAL CORD
20 pm
FIG. 5. (A) Calretinin-like immunoreactivity in laminae I and II of the spinal cord viewed in the sagittal plane. Many labeled fibers were seen running rostrocaudally. Arrow indicates the large cell body of a lamina I neuron. A recurrent process (arrowheads) was issued by a labeled neuron located at the bottom of lamina II. DAB reaction. Scale bar = 50 pm, X200. (B) Camera lucida drawings composed from two different sagittal sections showing two lamina II neurons with CR-LI. The top neuron resembles a lamina II stalked cell and the bottom neuron resembles an islet cell. Arrow indicates what may be an axon initial segment. The dotted lines indicate the borders of lamina II. Top = dorsal.
calcium binding protein is involved in distinct functional activities, although this generalization may be an oversimplification. Recently, Rausell and Jones [( 17). see also (9)] reported that calbindin D-28K and parvalbumin are present in two chemically distinct compartments in the thalamus. The small calbindin D28K-IR neurons of the thalamic VPM nucleus receive input from the caudal nucleus of the spinal trigeminal complex and project to superficial layers ofthe somatosensory cortex: the largeand medium-sized parvalbumin-IR neurons of the VPM nucleus receive input from the principal trigeminal nucleus and project to the middle layers of the somatosensory cortex. It was proposed that the pathway using calbindin was related to processing of nonlemniscal information, including pain and temperature sensations, although the route that contained parvalbumin was re-
lated to the trigeminal equivalent of the dorsal column lemniscal system ( 17). The laminar distribution of CR in the spinal cord suggests a role in sensory pathways, especially nociceptive pathways. The present study showed that the pyramidal tract and motor neurons in the ventral horn were not CR-IR. This is consistent with previous reports that motoneurons are usually devoid of CR ( I ,8.18.20). On the contrary, many CR-IR neurons were found in the superficial laminae of the dorsal horn, and laminae V, VII, VIII, and X. The densest staining, consisting of densely packed cell bodies and fiber networks, was found in lamina I and II, particularly lamina II. The CR-IR fibers in lamina II do not appear to be of primary afferent origin, because no reduction of CR-L1 in lamina II was apparent after dorsal root ganglio-
FIG. 6. The effects of multiple unilateral dorsal root ganghonectomies at the lumbar level (L25) on calretinin-like immunoreactivity in the L5 segment of the spinal cord. A and B are views taken from the same section. Comparing ipsilateral (PSI-) and contralateral (CONTRA-) sides, CR-IR fibers were largely reduced in the fasciculus gracilis (GR) on the ipsilateral side after ganghonectomy. However, CR-L1 appeared to increase in medial lamina II ipsilateral to ganghonectomy. Arrows in A and II point to the medial direction. DAB reaction, Scale bar (same for A and B) = 50 gm. X200.
CALRETININ
IN THE
RAT SPINAL
21
CORD
nectomy. It is noted, however, that the CR-L1 in lamina II may he regulated by primary afferent drive, as there appeared to be a slight increase in CR-IR cell bodies and dendrites in lamina II following ganglionectomy. This observation, however, requires further study to determine whether it is a consistent result of ganglionectomy. The specific distribution of CR-L1 in this region suggests a role for CR in spinal nociceptive transmission [for review, see (23)]. The locations of CR-IR neurons in the spinal gray matter in part matches that of the cells of origin of several ascending nociceptive tracts, i.e., spinothalamic. spinoreticular and spinomesencephalic tmcts (23). Recently, calbindin D-28K immunoreactive neurons in the spinal cord have been shown to project to supraspinal structures including the lateral reticular nucleus, parabrachial area. and periaqueductal grey (1 1.27). It remains to be determined whether some projection neurons in the spinal cord have CR-LI. It should be noted. however, that the large number of lamina II neurons containing CR-L1 are likely to be a mixed subpopulation of interneurons. Some of them (outer lamina 11 stalked cells) have been shown to respond to nociceptive input. and others (inner lamina II islet cells) respond exclusively to low-threshold mechanoreceptive input (2). Thus. although unlikely to be involved in motor control, CR may be involved in sensory processing of diverse modalities. In the present fasciculus gracilis immunoreactive
study, dense CR-IR fibers were found in the at the lumbosacral spinal cord, but very few fibers were localized in the fasciculus gracilis
at the rostra1 thoracic and cervical level. Loss of CR-L1 also was noticed in the fasciculus cuneatus at the Cl level. This phenomenon was not described in a previous report (I 8). It is known that dorsal column fasciculi consist of several fiber tracts of different origin [see (22) for a review]. These include the ascending
collaterals of primary afferents (direct dorsal column pathway). the axons of postsynaptic dorsal column neurons (second order dorsal column pathway), and the descending axons from the dorsal column nuclei. Many of the fibers that enter the dorsal columns do not reach the dorsal column nuclei (22). A portion of ascending fibers in the fasciculus gracilis has been shown to terminate on neurons in Clarke’s column (6). CR-L1 in the fasciculus gracilis is most likely of primary afferent origin because it was almost totally depleted by ganglionectomy at the lumbar level. The difference in CR-L1 in the fasciculus gracilis between the rostra1 and caudal spinal cord suggests that most CR-IR fibers in the fasciculus gracilis at the lumbosacral levels terminate below the level of the nucleus gracilis. Indeed, dense positive fiber networks were localized in Clarke’s column [( 18). the prcsent results] and a reduction in CR-L1 was found in Clarke’s column rostra1 to the level of a ganglionectomy. These results suggest that CR can be a useful neuronal marker for studying a central proprioceptive pathway. The mechanisms by which CR could be involved in the functional activity of neurons are still unclear. It is apparent, however, that the role of CR in cell function is not restricted to buffering the intracellular calcium concentration. In summary. the present results have provided a detailed mapping ofCR-containing neurons within the dorsal root ganglia and spinal cord. A unique distribution of cell bodies and fiber tracts were observed that suggests involvement of CR in noci-
ceptive and proprioceptive
functions.
AC‘KNOWLEDCEMEN-IS
We thank R. Dubner. viewing the manuscript.
D. J. Messersmith,
and R. 1.. Nahin for re-
REFERENCES 1. Arai. R.; Winsky. L.; Arai, M.; Jacobowitz. D. M. Immunohistochemical localization of calretinin in the rat hindbrain. J. Comp. Nemo]. 310:21-44; 1991. 2. Bennett. G. J.: Abdelmoumene, M.: Hayashi, H.; Dubner. R. Physiology and morphology of substantia gelatinosa neurons intracellularly stained with horseradish peroxidase. J. Comp. Neurol. 194: 809-827; 19x0. 3. Braun. K. Calcium-binding proteins in avian and mammalian central nervous system: Localization. development and possible functions. Prog. Histochem. Cytochem. 21: l-64: 1990. 4. Cervero, F.: Iggo. A. The substantia gelatinosa of the spinal cord: A critical review. Brain 103:7 17-772: 1980. 5. Christakos, S.: Gabrielides, C.; Rhoten. W. B. Vitamin D-dependent calcium binding proteins: Chemistry, distribution, functional considerations. and molecular biology. Endocr. Rev. 10:3-26; 1989. 6. Ganchrow, D.: Bernstein, J. J. Projections of caudal fasciculus gracilis to nucleus gracilis and other medullary structures. and Clarke’s nucleus in the rat. Brain Res. 205:383-390: 198 I, 7. Gobel. S. Golgi studies of the neurons in layer II of the dorsal horn ofthe medulla (trigeminal nucleus caudalis). J. Comp. Neural. 180: 395-414: 197x. 8. Jacobowitz. D. M.; Winsky. L. Immunocytochemical localization of calretinin in the forebrain of the rat. J. Comp. Neurol. 304:198218; 1991. 9. Jones, E. G.: Hendry, S. H. C. Differential calcium binding protein immunoreactivity distinguishes classes of relay neurons in monkey thalamic nuclei. Eur. J. Neurosci. 1:222-246: 1989. IO. Lee, K. H.: Chung. K.: Chung. J. M.: Coggeshall. R. E. Correlation of cell body size, axon size and signal conduction velocity for individually labelled dorsal root ganglion cells in the cat. J. Comp. Neurol. 243:335-346: 1986.
1 I. Menetrey, D.: de Pommery, J.: Thomasset. M.: Baimhridge. K. G. Calbindin-DZRK (CaBPZSK)-like immunoreactivity in ascending projections II. Spinal prqjections to brain stem and mesencephalic areas. Eur. J. Neurosci. 4:70-76; 1992. 12. Molander. C.: Xu. Q.; Grant, G. The cytoarchitectonic organization of the spinal cord in the rat. 1. The lower thoracic and lumbosacral cord. J. Comp. Neurol. 230:133-141: 1984. 13 Molander. C.: Xu. Q.; Rivero-Melian. C.: Grant. G. Cytoarchitectonic organization of the spinal cord in the rat: II. The cervical and upper thoracic cord. J. Comp. Neurol. 289:375-385: 1989. 14 Parmentier. M. Structure of the human cDNAs and genes coding for calbindin D28K and calretinin. In: Pochet. R.: Lawson, D. E. M.: Heizmann. C. W.. eds. Calcium binding proteins in normal and transformed cells. New York: Plenum Press: 1989:27-34. 15 Parmentier. M.: Lefort, A. Structure of the human brain calciumbinding protein calretinin and its expression in bacteria. Eur. J. Biochem. 196:79-85: 1991. 16. Paxinos. Cl.: Watson, C. The rat brain in stercotaxic coordinates. New York: Academic; 1986. 17. Rausell. E.; Jones. E. G. Chemically distinct compartments of the thalamic VPM nucleus in monkevs relav orincipal and spinal trigeminal pathways to different layers of td somatosensory cortex. J. Neurosci. I 1:226-237; 199 1, 18. Rtsibois. A.: Rogers. J. H. Calretinin in rat brain: An immunohistochemical study. Neuroscience 46: 10 l- 134: 1992. 19. Rogers, J. H. Calretinin: A gene for a novel calcium-binding protein expressed principally in neurons. J. Cell. Biol. 105: I343- 1353; 1987. 20. Rogers, J. H. Two calcium-binding proteins mark many chick sensory neurons. Neuroscience 3 1:697-709; 1989. 21. Rogers, J. H. Calretinin. In: Heizmann. C. W.. ed. Novel calcium-binding proteins. Berlin: Springer-Verlag; I99 I:25 l276.
22. Tracey, D. J. Ascending and descending pathways in the spinal cord. In: Paxinos. C.. ed. The rat nervous system. vol. 2. Hindbrain and spinal cord. Sydney: Academic Press: 1985:31 l-324. 23. Willis, W. D. The pain system. Base]: Karger: 1985. 24. Winsky. L.; Nakata, H.; Martin, B. M.; Jacobowitz. D. M. Isolation. partial amino acid sequence, and immunohistochemical localization of a brain-specific calcium-binding protein. Proc. Natl. Acad. Sci. USA 86:10139-10143; 1989. 25. Yamaguchi, T.; Winsky, L.; Jacobowitz, D. M. Calretinin, a neuronal calcium binding protein, inhibits phosphorylation of a 39 kDa syn-
aptic memebrane protein from rat brain cerebral cortex. Neurosci. Lett. 131:79-82; 1991. 26. Yamamoto, T.; Carr, P. A.: Baimbridge, K. G.: Nagy. J. I. Parvalbumin- and calbindin DZIK-immunoreactive neurons in the superficial layers of the spinal cord dorsal horn of rat. Brain Res. Bull 23:493-508; 1989. 27. Yoshida. S.; Senba, E.; Kubota, Y.; Hagihira, S.; Yoshiya. I.; Emson. P. C.: Tohyama, M. Calcium-binding proteins calbindin and parvalbumin in the superficial dorsal horn of the rat spinal cord. Neuroscience 37:839-848: 1990.
03hl-Y730/93 $6.00 + .oo Copyright c I993 Pergamon Press Ltd.
1993
Immunohistochemical Localization of Calretinin in the Dorsal Root Ganglion and Spinal Cord of the Rat K.
*Neurobiology Science,
REN.*’
M. A. RUDA*
AND
D. M. JACOBOWITZf
and Anesthesiology Branch, National Institute qf Dental Research and ilahoratoq~ ofClinical National Institute qj”A4ental Health, National Institutes of’ Health. Bethesda. MD 20892
Received
17 August
1992; Accepted
26 August
1992
REN. K.. M. A. RUDA AND D. M. JACOBOWITZ. Irnmllnohisto~hemic,ul loulimion of‘cdretinin uncl spind cord q/‘thc rut. BRAIN RES BULL 31( l/2) 13-22. 1993.-Calretinin (CR). a recently
in thr dord
root ~unglron
identified calcium-binding protein. is present in nervous tissue, including sensory pathways. where it may play an important role in regulation of cellular activity. Using immunocytochemistry, we examined the cellular localization of CR in dorsal root ganglia (DRG) and spinal cord of normal rats and after multiple unilateral dorsal root ganglionectomies. In DRG. CR-immunoreactive cell bodies and axons were a small subpopulation (10%) ofmediumto large-sized neurons. In the spinal cord. CR-like immunoreactivity (LI) in neurons and fibers was found in all laminae except motoneurons. Dense fiber networks were also found in Clarke’s column. The densest staining of both cell bodies and fibers was in the superficial laminae. especially lamina II. and in the lateral spinal and lateral cervical nuclei. CR-immunoreactive fibers were also observed in the fasciculi cuneatus and gracilis. Fasciculus gracilis exhibited the greatest number of labeled axons at the lumbosacral levels, but few labeled axons were found at the rostra1 thoracic and cervical levels. In contrast. the corticospinal tract at the base of the dorsal column was devoid of CR-immunoreactive fibers. Unilateral multiple lumbar ganglionectomies resulted in a loss of CR-L1 in the dorsal columns ipsilateral to the surgery. In the spmal gray matter ipsilateral to the ganglionectomies. CR-L1 was reduced in Clarke’s column and slightly enhanced in the medial third of lamina II. Our observations demonstrate a unique distribution pattern of CR-L1 compared to other calcium-binding proteins in the spinal cord. and suggest a role for CR in nociceptive and proprioceptive pathways. Spinal cord Calretinin Calcium-binding proteins
Dorsal root ganglion Rat
Dorsal column
(CR) is a newly discovered member of the family of calcium-binding proteins (I 9.24). It is closely related to another calcium-binding protein, calbindin D-28K (3J.14). The primary translation product of CR mRNA is 29 kDa (19). CR and calbindin D-28K have a 50-60s cDNA and amino acid sequence homology ( 14.15.19.24). RNA blots have detected CR mRNA in the brain but not in peripheral tissues ( 15,I9). Thus, CR appears to he primarily a neuronal protein. CR contains six potential calcium binding domains, five of which are thought to be active in binding calcium (14). Although it seems obvious that CR may modulate calcium concentrations in cells, its functional significance has not been determined. It is thought that CR, as well as some other calcium-binding proteins, perform unknown functions other than just buffering cellular calcium concentrations (2 I). This view is supported by the specific and differential distribution patterns of calcium-binding proteins in the nervous system (1.8.17-19,26,27). It has been
should
be addressed
Immunocytochemistry
recently shown that CR is involved in the regulation of protein phosphorylation in the brain (25). Although the distribution of CR immunoreactivity in the rat and chick brain has been studied in great detail. a detailed account of the distribution in the spinal cord and dorsal root ganglia has not been reported (1,8,18,20). Using a specific antiserum raised against guinea pig CR. originally named protein 10 (24). a detailed description of CR-like immunoreactivity (LI) in the cervical through lumbosacral spinal cord is presented in this report. We also describe the localization of CR-L1 in a subpopulation of dorsal root ganglion (DRG) neurons. The effects of unilateral multiple ganglionectomies on CR-L1 in the spinal cord were also studied.
CALRETININ
’ Requests for reprints 20892.
Ganglionectomy
METHOD
Ten male Sprague-Dawley rats (200-300 the present study. Five rats received unilateral
to K. Ren. Ph.D. NAB/NlDR/NIH
13
Building
30. Room
B-20. 9000 Rockville
g) were used in multiple dorsal
Pike. Bethesda,
MD
14
root gan~ionectomies under sterile and anesthetic conditions and were allowed to survive one (n = 2) to two (n = 3) weeks after the surgery. All rats were deeply anesthetized with sodium pentobarbital and perfused transcardially with 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. The extent of the ganglionectomies was verified after the perfusion. The cervical, thoracic, lumbar and sacral spinal cord and DRG were removed, immersed in the same fixative overnight at 4”C, and transferred to 30% sucrose (w/v) in phosphate buffer for several days for cryoprotection. Thirty micron-thick sections were cut with a cryostat at -20°C in either the transverse or sagittal plane. The sections were collected in phosphate buffered saline (PBS) and processed for immunocytochemistry. The primary antiserum against guinea pig CR was raised in rabbits as described (24). The specificity of the antiserum has been examined previously (1,8). Incubation of CR antiserum with purified CR eliminated immunostaining of the brain sections (1,s). Free-floating tissue sections were rinsed in PBS with 0.75% Triton X-100 for I h, 5%) normal goat serum (NGS) in PBS for 30 mitt, then incubated with anticalretinin antiserum ( I :7500- 10,000 in PBS) with I-3% NGS for 48-72 h with gentle agitation. After two washes in PBS (10 min each), and a 20-min rinse in PBS with 5% NGS, the sections were incubated with biotinylated goat antirabbit IgG (Vector, I:200 in PBS) for 30-60 min. Following rinses in PBS (10 min X 2) and PBS with 5% NGS (20 min), the sections were then incubated with avidin and biotinylated horseradish peroxidase complex (Vector, 1:50 in PBS) for 30-60 min. After washes in PBS or Tris HCl-buffered saline (TBS), the tissue sections were reacted with 0.05% diaminobenzadine dihydrochloride (DAB, Sigma) in 0.1 M phosphate buffer containing 0.003% hydrogen peroxide for 6 min. In some experiments, the sections were preincubated in 0.2% nickel ammonium sulfate in 0.1 M phosphate buffer for 10 min before incubation with DAB. Both procedures produced the same immunostaining pattern in the tissue although the nickel ammonium sulfate incubation resulted in increased contrast. Some sections were later counterstained with cresyl violet to allow visualization of the general histology of the tissue. Control sections were processed with the same method except that the primary antiserum was omitted. The sections were finally rinsed in PBS, mounted on gelatin-coated slides, dehydrated in alcohols, cleared in xylenes, and coverslipped with Eukitt (Kindler, Germany). Immunostaining intensity and packing density of CR-immunoreactive (IR) cell bodies and fibers were subjectively assessed as described (1,8). The morphology of some spinal neurons was traced with a camera lucida. CR-L1 also was analyzed with a computer-assisted image analysis system (Image 1.41, National Institutes of Health). The size of labeled neurons was determined by tracing the outlines of cell bodies and calculating their cross-sectional areas (pm’). An effort was made to measure neurons with a visibfe nucleus and nucleolus. When nucleolus was indistinguishable, especially for small neurons, neurons with clear nuclei were measured. Spinal neurons (n = 1736) were measured in randomly selected sections from five animals and arbitrarily divided into three groups according to their sizes; that is, small, < 150 pm2; medium, 150-300 pm*; and large, > 300 Frn’ ( 1). DRG neurons were usually larger and thus grouped differently: small, < 1000 Frn’; medium, 1000-2000 pm*: and large, > 2000 pm* (10). Rat spinal cord structures were identified and named according to the stereotaxic atlases of Paxinos and Watson ( 16) and Molander et al. f 12.13).
RI-N. RL;DA /ihI)
JAC‘OROWI I /
RtSl.II.IS
The CR-L1 was observed at all segmental levels studied. Some segmental differences, especially in specialized nuclear groups. were found. For instance, the CR-IR cell bodies at the lumbosacral spinal cord were more abundant than those at the ccrvicothoracic spinal cord. The densest immunoreactive staining was found in the superficial dorsal horn, especially lamina II. and the lateral spinal and lateral cervical nuclei. Examples of the CR-L1 in the representative segments of the spinal cord are illustrated in Fig. I. An analysis of the CR-IR cell bodies and fibers is summarized in Table 1. A few examples of CR-IR neurons are shown in Fig. 2. Many nuclei were stained denser than the cytoplasm (Fig. 2C, D). The nucleolus was not labeled (Fig. 2A, E). Glia cells were not stained. Dorsul root ganglion. Approximately 10% of total number of neurons in the DRG were labeled with CR antiserum. No segmental difference was found at cervical through sacral levels. Variable staining density, not correlated with cell size, was found (Fig. 3). Cross-sectional area of the labeled neurons ranged from 300 to over 3000 pm*, although much of the cells were medium to large sized (90%). Labeled axons, especially with large diameters (>2 gm). were found in the ganglia and dorsal roots. Axons originating from cell bodies could be easily seen (Fig. 3B). The CR-IR axons could be traced into the dorsal root entry zone of the spinal cord (Fig. 4). ~F?z~nu I. The largest lamina I cells, very likely co~esponding to Waldeyer cells or marginal ceils [see (23), for a review], were labeled. In transverse sections, Waldeyer cells were seen under Lissauer’s tract or in the middle third of lamina I. They typically have a mediolaterally oriented fusiform cell body with dendrites extending along the border of lamina I (Fig. 2B). The crosssectional areas of the Waldeyer cell bodies were 100-300 mm*,two to three times larger than lamina II CR-IR neurons. In addition to Waldeyer cells. smaller round cell bodies were also CR-IR. A CR-IR plexus of nerve fibers was found in lamina I. However, it did not appear to be quite as dense as that in lamina II. L~r7ziF~uII. A densely stained band, corresponding to lamina II in the superficial dorsal horn, was found in all segments of the cord examined (Fig. 1). The immunor~ctive materials consisted of high density neuropil and densely packed cell bodies (Figs. 2B and 4). The fiber plexus staining intensity was approximately the same for outer and inner portions of lamina II through most spinal segments, However, CR-IR ceil bodies were densely packed along the border of laminae II and III in the rostra1 cervical cord. At different segment levels, CR exhibited variable density between the lateral and medial aspects of lamina II. At cervical levels, lateral lamina II was typically more densely stained than medial lamina II. At lumbar levels, the pattern was similar. although CR-L1 was generally denser than at cervical levels. In Contras& at both thoracic and sacral levels, CR-L1 was comparatively more dense from the lateral to the medial part of Iamina II (Fig. I). Cross-sectional areas of most Iamina II CR-IR neurons were smaller than 100 ym2. In the transverse plane, the ceil bodies appeared as round or oval in shape and dendritic orientations were not clear, In sagittal sections, however, many lamina II CR-IR dendrites were observed extending rostrocaudahy (Figs. 1E and SA). One type of lamina II CR-IR neuron, mainly found in outer lamina II, had ventrally directed dendrites (Fig. SB), resembling those of the limiting cell or stalked cell described previously [(2,7), also see (4), for a review]. The other type of CR-IR neuron located in both the outer and inner portion of lamina II resembled the central cell or islet cell [(2,7), also see (4). for a review], whose
CAL.RETININ
IN THE RAT SPINAL CORD
FIG. 1. The dist~bution of cairetinin-like iummunoreacfivity in the cervical C5 (A). lumbar L5 (B), thoracic T3 (C). and sacral S2 (If) spinal cord. The densest staining was found in the superficial dorsal horn and the lateral spinal nucleus. Note that the immunoreactive bands in the superficiai dorsal horn are less dense medially at the lumbar and cervical levels. Also note the differential staining pattern in the dorsal columns at cervical, thoracic, lumbar, and sacral levels. Laminae 111and IV. pyramidal tract (PT), and ventral root (e.g., arrow heads in A) generally lack CR-LI. A sagittai section from the lumbar enlargement, cut through a sagittal plane indicated by an arrow in B, is shown in E. CC, Clarke’s column; CU, fasciculus cuneatus; DC, dorsal column; DGC, dorsal gray commissure; GR, fasciculus gracilis; IML, intermediolateral nucieus; LSN, lateral spinal nucleus. Roman numerals I-X indicate spinal cord laminae. (A-D) Nickel enhanced DAB reaction; (E) DAB reaction. Scale bar (same for A-E) = 200 pm, X37.5.
dendrites mainly travel rostrocaudally. Occasionally, the recurrent process on these neurons was Seen (Fig. 5A). Laminuc III und IV. This area of the spinal cord gray matter consistently exhibited the least CR-LI. Only few scattered cell bodies and axons were labeled. The neuropil in this region was generally devoid of CR-LI. Laminae L’ and W. Dense immunoreactive neuropil and numerous labeled cells were found in these laminae. Most immunoreactive neurons were medium to large multipolar neurons
(Fig. I B. E). The immunoreactive
reaction product was evenly distributed in the cytoplasm. The dendrites and axons could be traced up to 200-400 pm away from the cell bodies. An example of a lamina VI multipolar neuron is shown in Fig. 2A. Pyramidal and smail bipolar neurons also were seen. Laminae V-VI neurons with CR-LI were more abundant at the lumbar level than at cervical, thoracic, and sacral levels. At the lumbar level, immunoreactive cell bodies and neuropil were mainly distributed in the middle third of the iaminae (Fig. 1B).
IABLE ~MMUNOHIST~HE~ICAL
LOCALIZATION ---~-
1
OF CALRETININ
IN THE RAT SPINAL
(‘ell Bodies Spinal Cord Region
Lamina I Lamina II Laminae III/IV Lamina V/VI Cervicothoracic Lumbo~~ral Lamina VII Cervicothoraeic Lumbosacral Lamina VIII Lamina IX Lamina X Clarke’s column Lissaeur’s tract Lateral spinal nucleus Gracile fasciculus LumbosacraI Cervicothoracic Cuneate faseiculus Pyramidal tract
Shape*
Size?
Staimng Intensityt
CORD
__...Packing Densityg
I-ibcn __
Packing Density*
0, f
S. M
t, ++
0. f 0. f, m
s s
++ +. ++
f. P, m f. p. m
S. M, l..
+. ++
+
ti
S. M, L
f. if
C-F, +++
++, +++
f, m f, m f. m
M. L M. L M. 1.
+, ++ +. ++ +. ++
+
+. ++
t, t+ t
0. f
S. M
+, ++
-t -
0, f
S
f, ++
+-, +
++. t ++ i t, t-i + +t. i-i+ i tt -I-
if,
ir -t-t+
-
i -! r
+ f. i +-t. ft fS-. i-i+,
ft
* o, oval or round; f, fusiform; p, pyramidal; m, muIti~Iar. i S, < I50 pm’: M, 150-300 pm’; L, >3OO pm’; (rm2, cell crosssectional area). $ +t-, intense; +, light. 9:+++, high density; ++, moderate density; t, low density; +, rarely found; -, not found.
# +++, high-density neuropil; ++, moderate-density neuropil; +, low-density neuropil; -. not found; ft. fiber tract.
Laminae Hi and VIII. Many medium- to large-size neurons with a variety of shapes were stained. Multipolar neurons were the most abundant in this intermediate region of the spinal cord gray matter. The cross-sectional area ranged from 100 to i000 pm’. Their dendrites were inte~in~~, and most of them apneared to be oriented do~oventrally when viewed in the sagittal plane (Fig. I E). Lamina IX. Although CR-L1 was present in fiber networks, the large motoneurons in lamina IX were not labeled by CR antiserum. This was clearly shown in sections counterstained with cresyl violet (data not shown). Consistently, the ventral roots originating from motoneuron pools were not labeled (Fig. IA). Lamina X. CR-IR cell bodies were frequently found dorsal to the central canal. Occasionally, CR-IR cell bodies were found lateral to the central canal, but essentially no labeled neurons were found ventral to the central canal. Most labeled neurons were small. A few medium flattened cells with cell bodies oriented m~olaterally were seen in this region. Densely packed immunoreactive cell bodies were clustered in the dorsal gray commissure (DGC) of the sacral spinal cord. Their processes were distributed horizontally, reaching the intermediolateral nucleus. Ceil bodies in the DGC also sent processes running dorsally towards the edge ofthe gray matter. A typical section illustrating CR-L1 in the DGC is shown in Fig. ID. Other special spinal nuclei. At the upper cervical spinal cord (Cl-3) immunoreactive fiber networks were aggregated in the territories of the internal basilar nucleus and central cervical nucleus. Clarke’s column was clearly distinguishable at the thoracic level by a densely labeled neuropil (Fig. 1C). Very few CR-
IR cell bodies were found in the central cervical nucleus, Clarke’s column, intermediate medial nucleus, and intermediate lateral nucleus at the related segments. White matter. The white matter surrounding the spinal cord contained numerous CR-IR axons. The dorsal lateraf fascicufus was densely stained. A few labeled cell bodies were occasionally seen in the Iateral spinal nucleus and lateral cervical nucleus. In some sections, elongated neurons with dendritic processes extending from the lateral spinal nucleus to the dorsal horn could be seen. In Lissauer’s tract, few fine fibers were stained (Fig. 4). Many fibers in the dorsal column fasciculi cuneatus and gracilis were immunoreactive to the calretinin antiserum. Interestingly, although clear immunostaining was found in the fasciculus gracilis at the Iumbosacral levels, very few immunoreactive fibers were observed in the fasciculus gracihs at the rostra1 thoracic and cervical levels (Fig. I). The fasciculus cuneatus was usually full of immunoreactive fibers, but at the CI level, lateral part of the fasciculus cuneatus dorsal to the medial portion ofthe dorsal horn also was not stained. Most parts of the ventrat and lateral white matter were stained by CR antiserum. The pyramidal tract in the ventromedial portion of the dorsal coiumn and fibers in the dorsal midline area (tail zone) were not CR-IR. Efects of UnilateralMultiple Dorsal Root Ganglionectomies Five rats received unilateral multiple ganglionectomies primarily at the lumbar levels. The locations ofthe ganglionectomies were L2-5, L I-5, L3-S I, and L2-6. The other one rat had L I 2 ganghonectomies and L3 and L6 dorsal root axotomies. The ganghonectomized rats were perfused 1 (n = 2) or 2 (n = 3) weeks after the surgery.
CALRETININ
IN THE
RAT SPINAL
17
CORD
FIG. ?. Photomicrographs illustrating calretinin-immunoreactive cell bodies in the spinal cord. (A) A multipolar neuron (arrow) in lamina Vi. Note uniform immunostaining in the cytoplasm, nucleus, and processes. The nucleolus, the clear dot in the center of the cell. was not stained. Arrowheads indicate an axon-like process. (B) A Waldeyer type lamina I neuron (arrow) in the sacral spinal cord with mediolateraily extended dendrites (arrowheads). Another large round lamina I neuron (asterisk) is also present in this section. Numerous CR-IR cell bodies are also found in lamina II. (C) A multipolar CR-IR neuron (arrow) in lamina VII. Note that the nucleus is stained more densely than the cytoplasm. (D) A pyramidal neuron (arrow) in lamina V. The nucleus is densely stained. (E) Arrow indicates a bipolar CR-IR neuron in lamina VIII. Nickel enhanced DAB reaction. Scale bar (same for 4-E) : 25 pm. X316
There was a loss of CR-L1 in the dorsal column fascicuius gracilis at related segments ipsilateral to the ganglionectomy or axotomy. L6 dorsal root axotomy with L4, 5 DRG intact resulted in a loss of CR-LI in the dorsoiaterai corner of the ipsilateral dorsal column at the related segment. L2-5 ganghonectomies essentially led to a total loss of immunostaining in the ipsiiaterai dorsal column (Fig. 6). The changes in spinal gray matter were less apparent after gangiionectomy. CR-L1 was reduced slightly in the nucleus dorsalis (Clark’s column) rostrai to the surgery levels. The CR-L1 was slightly enhanced in the medial third of lamina 11 at the surgical level ipsiiaterai to the gangiionectomy. In four animals with at least L4, 5 gangiionectomies, the staining intensity and packing density of CR-IR cell bodies was slightly increased especially in the medial end of iamina II (Fig. 6).
DISCUSSION A distinct and reproducible CR-L1 was observed in the cervical through sacral DRG and spinal cord of the rat in the present study. The CR-L1 was extensively distributed in spinal cord structures including iaminae I, 11, V-VIII, and X, the lateral spinal and lateral cervical nuclei, and the dorsal column fascicuii cuneatus and gracilis. The CR-L1 was found in the neuronal cytoplasm. nucleus. dendrites, and axons. The immunocytochemical localization of CR in the spinal cord has been briefly described by Rogers (20) and Resibois and Rogers (18). The present results are generally consistent with their results. It was further found that the dorsal root ganglion~tomy primarily resuited in a loss of CR-L1 in the dorsal column. The CR-L1 in spinal gray matter was largely preserved after the gangiionectomy. Obviously. although CR-IR primary afferents project through
FIG. 3. Dorsal root ganglion neurons stained with calretinin antiserum. (A) The neurons exhibit both dense (arrows) and moderate (arrowheads) cytoplasmic immunoreactivity. Most CR-IR cells are medium to large sized. (B) Arrowheads point to a CR-IR axon as it exits from the neuronal cell body. DAB reaction. Scale bars: A. 100 pm, X100: B. 25 pm. x380.
the spinal cord, few CR-IR primary afferents terminating in the spinal gray matter, and spinal neurons exhibit intrinsic CR-LI. The differential distribution patterns of different calciumbinding proteins are most impressive. It is apparent from this study that CR is differentially distributed in the spinal cord. when compared with the locakation of other calcium-binding proteins. CR was most abundandy distributed in iamina II and the lateral spinal and lateral cervical nuclei of the spinal cord. The CR-L1 in laminae III, IV, and IX was either very light or absent. In contrast, parvalbumin is primarily localized in inner lamina II and outer lamina III of the spinal cord (26.27). A close
relative of CR, calbindin D-28K, is most abundant in laminae 1and II. Its distribution pattern is homogeneous through different spinal segments as compared to CR (27). In contrast to CR, caimodulin-LI was not detectable in lamina I1 of the dorsal horn, but did exist in the motoneuron pools ofthe ventral horn (lamina IX) (unpublished observations). The specific and differential localization of calcjum-binding proteins has also been noted in the brain [for examples, ( 1,8,1Xl]. The functional significance of differential or complementary localization of different calcium binding proteins in the nervous system is not understood. It is reasonable to suggest that each
FIG. 4. Calretinin-like immunoreactivity in the dorsal root entry zone in a transverse section from CS spinal cord. Thick lab&d fibers can be traced into the dorsal column (arrowheads) from the dorsal root. Few fine labeled fibers are seen in Lissauer’s tract (LISS). DR, dorsal root: CU. fasciculus cuneatus. Nickel enhanced DAB reaction. Scale bar = 50 pm* X200.
(‘ALRETININ
I9
IN THE RAT SPINAL CORD
20 pm
FIG. 5. (A) Calretinin-like immunoreactivity in laminae I and II of the spinal cord viewed in the sagittal plane. Many labeled fibers were seen running rostrocaudally. Arrow indicates the large cell body of a lamina I neuron. A recurrent process (arrowheads) was issued by a labeled neuron located at the bottom of lamina II. DAB reaction. Scale bar = 50 pm, X200. (B) Camera lucida drawings composed from two different sagittal sections showing two lamina II neurons with CR-LI. The top neuron resembles a lamina II stalked cell and the bottom neuron resembles an islet cell. Arrow indicates what may be an axon initial segment. The dotted lines indicate the borders of lamina II. Top = dorsal.
calcium binding protein is involved in distinct functional activities, although this generalization may be an oversimplification. Recently, Rausell and Jones [( 17). see also (9)] reported that calbindin D-28K and parvalbumin are present in two chemically distinct compartments in the thalamus. The small calbindin D28K-IR neurons of the thalamic VPM nucleus receive input from the caudal nucleus of the spinal trigeminal complex and project to superficial layers ofthe somatosensory cortex: the largeand medium-sized parvalbumin-IR neurons of the VPM nucleus receive input from the principal trigeminal nucleus and project to the middle layers of the somatosensory cortex. It was proposed that the pathway using calbindin was related to processing of nonlemniscal information, including pain and temperature sensations, although the route that contained parvalbumin was re-
lated to the trigeminal equivalent of the dorsal column lemniscal system ( 17). The laminar distribution of CR in the spinal cord suggests a role in sensory pathways, especially nociceptive pathways. The present study showed that the pyramidal tract and motor neurons in the ventral horn were not CR-IR. This is consistent with previous reports that motoneurons are usually devoid of CR ( I ,8.18.20). On the contrary, many CR-IR neurons were found in the superficial laminae of the dorsal horn, and laminae V, VII, VIII, and X. The densest staining, consisting of densely packed cell bodies and fiber networks, was found in lamina I and II, particularly lamina II. The CR-IR fibers in lamina II do not appear to be of primary afferent origin, because no reduction of CR-L1 in lamina II was apparent after dorsal root ganglio-
FIG. 6. The effects of multiple unilateral dorsal root ganghonectomies at the lumbar level (L25) on calretinin-like immunoreactivity in the L5 segment of the spinal cord. A and B are views taken from the same section. Comparing ipsilateral (PSI-) and contralateral (CONTRA-) sides, CR-IR fibers were largely reduced in the fasciculus gracilis (GR) on the ipsilateral side after ganghonectomy. However, CR-L1 appeared to increase in medial lamina II ipsilateral to ganghonectomy. Arrows in A and II point to the medial direction. DAB reaction, Scale bar (same for A and B) = 50 gm. X200.
CALRETININ
IN THE
RAT SPINAL
21
CORD
nectomy. It is noted, however, that the CR-L1 in lamina II may he regulated by primary afferent drive, as there appeared to be a slight increase in CR-IR cell bodies and dendrites in lamina II following ganglionectomy. This observation, however, requires further study to determine whether it is a consistent result of ganglionectomy. The specific distribution of CR-L1 in this region suggests a role for CR in spinal nociceptive transmission [for review, see (23)]. The locations of CR-IR neurons in the spinal gray matter in part matches that of the cells of origin of several ascending nociceptive tracts, i.e., spinothalamic. spinoreticular and spinomesencephalic tmcts (23). Recently, calbindin D-28K immunoreactive neurons in the spinal cord have been shown to project to supraspinal structures including the lateral reticular nucleus, parabrachial area. and periaqueductal grey (1 1.27). It remains to be determined whether some projection neurons in the spinal cord have CR-LI. It should be noted. however, that the large number of lamina II neurons containing CR-L1 are likely to be a mixed subpopulation of interneurons. Some of them (outer lamina 11 stalked cells) have been shown to respond to nociceptive input. and others (inner lamina II islet cells) respond exclusively to low-threshold mechanoreceptive input (2). Thus. although unlikely to be involved in motor control, CR may be involved in sensory processing of diverse modalities. In the present fasciculus gracilis immunoreactive
study, dense CR-IR fibers were found in the at the lumbosacral spinal cord, but very few fibers were localized in the fasciculus gracilis
at the rostra1 thoracic and cervical level. Loss of CR-L1 also was noticed in the fasciculus cuneatus at the Cl level. This phenomenon was not described in a previous report (I 8). It is known that dorsal column fasciculi consist of several fiber tracts of different origin [see (22) for a review]. These include the ascending
collaterals of primary afferents (direct dorsal column pathway). the axons of postsynaptic dorsal column neurons (second order dorsal column pathway), and the descending axons from the dorsal column nuclei. Many of the fibers that enter the dorsal columns do not reach the dorsal column nuclei (22). A portion of ascending fibers in the fasciculus gracilis has been shown to terminate on neurons in Clarke’s column (6). CR-L1 in the fasciculus gracilis is most likely of primary afferent origin because it was almost totally depleted by ganglionectomy at the lumbar level. The difference in CR-L1 in the fasciculus gracilis between the rostra1 and caudal spinal cord suggests that most CR-IR fibers in the fasciculus gracilis at the lumbosacral levels terminate below the level of the nucleus gracilis. Indeed, dense positive fiber networks were localized in Clarke’s column [( 18). the prcsent results] and a reduction in CR-L1 was found in Clarke’s column rostra1 to the level of a ganglionectomy. These results suggest that CR can be a useful neuronal marker for studying a central proprioceptive pathway. The mechanisms by which CR could be involved in the functional activity of neurons are still unclear. It is apparent, however, that the role of CR in cell function is not restricted to buffering the intracellular calcium concentration. In summary. the present results have provided a detailed mapping ofCR-containing neurons within the dorsal root ganglia and spinal cord. A unique distribution of cell bodies and fiber tracts were observed that suggests involvement of CR in noci-
ceptive and proprioceptive
functions.
AC‘KNOWLEDCEMEN-IS
We thank R. Dubner. viewing the manuscript.
D. J. Messersmith,
and R. 1.. Nahin for re-
REFERENCES 1. Arai. R.; Winsky. L.; Arai, M.; Jacobowitz. D. M. Immunohistochemical localization of calretinin in the rat hindbrain. J. Comp. Nemo]. 310:21-44; 1991. 2. Bennett. G. J.: Abdelmoumene, M.: Hayashi, H.; Dubner. R. Physiology and morphology of substantia gelatinosa neurons intracellularly stained with horseradish peroxidase. J. Comp. Neurol. 194: 809-827; 19x0. 3. Braun. K. Calcium-binding proteins in avian and mammalian central nervous system: Localization. development and possible functions. Prog. Histochem. Cytochem. 21: l-64: 1990. 4. Cervero, F.: Iggo. A. The substantia gelatinosa of the spinal cord: A critical review. Brain 103:7 17-772: 1980. 5. Christakos, S.: Gabrielides, C.; Rhoten. W. B. Vitamin D-dependent calcium binding proteins: Chemistry, distribution, functional considerations. and molecular biology. Endocr. Rev. 10:3-26; 1989. 6. Ganchrow, D.: Bernstein, J. J. Projections of caudal fasciculus gracilis to nucleus gracilis and other medullary structures. and Clarke’s nucleus in the rat. Brain Res. 205:383-390: 198 I, 7. Gobel. S. Golgi studies of the neurons in layer II of the dorsal horn ofthe medulla (trigeminal nucleus caudalis). J. Comp. Neural. 180: 395-414: 197x. 8. Jacobowitz. D. M.; Winsky. L. Immunocytochemical localization of calretinin in the forebrain of the rat. J. Comp. Neurol. 304:198218; 1991. 9. Jones, E. G.: Hendry, S. H. C. Differential calcium binding protein immunoreactivity distinguishes classes of relay neurons in monkey thalamic nuclei. Eur. J. Neurosci. 1:222-246: 1989. IO. Lee, K. H.: Chung. K.: Chung. J. M.: Coggeshall. R. E. Correlation of cell body size, axon size and signal conduction velocity for individually labelled dorsal root ganglion cells in the cat. J. Comp. Neurol. 243:335-346: 1986.
1 I. Menetrey, D.: de Pommery, J.: Thomasset. M.: Baimhridge. K. G. Calbindin-DZRK (CaBPZSK)-like immunoreactivity in ascending projections II. Spinal prqjections to brain stem and mesencephalic areas. Eur. J. Neurosci. 4:70-76; 1992. 12. Molander. C.: Xu. Q.; Grant, G. The cytoarchitectonic organization of the spinal cord in the rat. 1. The lower thoracic and lumbosacral cord. J. Comp. Neurol. 230:133-141: 1984. 13 Molander. C.: Xu. Q.; Rivero-Melian. C.: Grant. G. Cytoarchitectonic organization of the spinal cord in the rat: II. The cervical and upper thoracic cord. J. Comp. Neurol. 289:375-385: 1989. 14 Parmentier. M. Structure of the human cDNAs and genes coding for calbindin D28K and calretinin. In: Pochet. R.: Lawson, D. E. M.: Heizmann. C. W.. eds. Calcium binding proteins in normal and transformed cells. New York: Plenum Press: 1989:27-34. 15 Parmentier. M.: Lefort, A. Structure of the human brain calciumbinding protein calretinin and its expression in bacteria. Eur. J. Biochem. 196:79-85: 1991. 16. Paxinos. Cl.: Watson, C. The rat brain in stercotaxic coordinates. New York: Academic; 1986. 17. Rausell. E.; Jones. E. G. Chemically distinct compartments of the thalamic VPM nucleus in monkevs relav orincipal and spinal trigeminal pathways to different layers of td somatosensory cortex. J. Neurosci. I 1:226-237; 199 1, 18. Rtsibois. A.: Rogers. J. H. Calretinin in rat brain: An immunohistochemical study. Neuroscience 46: 10 l- 134: 1992. 19. Rogers, J. H. Calretinin: A gene for a novel calcium-binding protein expressed principally in neurons. J. Cell. Biol. 105: I343- 1353; 1987. 20. Rogers, J. H. Two calcium-binding proteins mark many chick sensory neurons. Neuroscience 3 1:697-709; 1989. 21. Rogers, J. H. Calretinin. In: Heizmann. C. W.. ed. Novel calcium-binding proteins. Berlin: Springer-Verlag; I99 I:25 l276.
22. Tracey, D. J. Ascending and descending pathways in the spinal cord. In: Paxinos. C.. ed. The rat nervous system. vol. 2. Hindbrain and spinal cord. Sydney: Academic Press: 1985:31 l-324. 23. Willis, W. D. The pain system. Base]: Karger: 1985. 24. Winsky. L.; Nakata, H.; Martin, B. M.; Jacobowitz. D. M. Isolation. partial amino acid sequence, and immunohistochemical localization of a brain-specific calcium-binding protein. Proc. Natl. Acad. Sci. USA 86:10139-10143; 1989. 25. Yamaguchi, T.; Winsky, L.; Jacobowitz, D. M. Calretinin, a neuronal calcium binding protein, inhibits phosphorylation of a 39 kDa syn-
aptic memebrane protein from rat brain cerebral cortex. Neurosci. Lett. 131:79-82; 1991. 26. Yamamoto, T.; Carr, P. A.: Baimbridge, K. G.: Nagy. J. I. Parvalbumin- and calbindin DZIK-immunoreactive neurons in the superficial layers of the spinal cord dorsal horn of rat. Brain Res. Bull 23:493-508; 1989. 27. Yoshida. S.; Senba, E.; Kubota, Y.; Hagihira, S.; Yoshiya. I.; Emson. P. C.: Tohyama, M. Calcium-binding proteins calbindin and parvalbumin in the superficial dorsal horn of the rat spinal cord. Neuroscience 37:839-848: 1990.