Commissural propriospinal connections between the lateral aspects of laminae III-IV in the lumbar spinal cord of rats.

Commissural Propriospinal Connections
between the Lateral Aspects of Laminae
III–IV in the Lumbar Spinal Cord of Rats
MIHA´LY PETKO´,1 GA´BOR VERESS,1 GYO¨RGY VEREB,2 JON STORM-MATHISEN,3
AND MIKLO´S ANTAL1*
1Department of Anatomy, Histology and Embryology, Faculty of Medicine, Medical and
Health Science Center, University of Debrecen, H-4012 Debrecen, Hungary
2Department of Biophysics and Cell Biology, Faculty of Medicine, Medical and Health
Science Center, University of Debrecen, H-4012 Debrecen, Hungary
3Anatomical Institute, University of Oslo, N-0317 Oslo, Norway
ABSTRACT
It has been established that there is a strong functional link between sensory neural circuits
on the two sides of the spinal cord. In one of our recent studies we provided a morphological
confirmation of this functional phenomenon, presenting evidence for the presence of a direct
commissural connection between the lateral aspects of the dorsal horn on the two sides of the
lumbar spinal cord. By using a combination of neural tracing and immunocytochemical detection
of neural markers like vesicular glutamate transporters, glutamic acid decarboxylase, glycine
transporter, and met-enkephalin (which are characteristic of various subsets of excitatory and
inhibitory neurons), we investigated here the distribution, synaptic relations, and neurochemical
characteristics of the commissural axon terminals. We found that the cells of origin of commis-
sural fibers in the lateral aspect of the dorsal horn were confined to laminae III–IV and projected
to the corresponding area of the contralateral gray matter. Most of the commissural axon
terminals established synaptic contacts with dendrites. Axospinous or axosomatic synaptic con-
tacts were found in limited numbers. We demonstrated that interactions among commissural
neurons also exist. More than three-fourths of the labeled axon terminals were immunostained
for glutamic acid decarboxylase and/or glycine transporter, but none of them showed positive
immunoreaction for met-enkephalin and vesicular glutamate transporters. The results indicate
that there is a substantial reciprocal commissural synaptic interaction between the lateral
aspects of laminae III–IV on the two sides of the lumbar spinal cord and that this pathway may
transmit both inhibitory and excitatory signals to their postsynaptic targets. J. Comp. Neurol.
480:364–377, 2004. © 2004 Wiley-Liss, Inc.
Indexing terms: spinal dorsal horn; propriospinal neurons; neural tracing; immunocytochemistry;
amino acid neurotransmitters
It has generally been accepted that there is a strong
functional link between neural circuits underlying sen-
sory information-processing mechanisms on the two sides
of the spinal cord. It has been demonstrated that experi-
mentally induced unilateral hind paw inflammation pro-
duces bilateral changes in the expression of intermediate-
early genes (Williams et al., 1990; Herdegen et al.,
1991a,b; Ren and Ruda, 1996), substance P, calcitonin
gene-related peptide (CGRP), glutamic acid decarboxylase
(GAD) immunoreactivity (Sluka et al., 1992; Mapp et al.,
1993), and NADPH-diaphorase activity (Solodkin et al.,
1992) in the superficial laminae of the dorsal horn. After
mechanical or thermal noxious stimulation of the hind
limb and tail of decerebrated spinal rats, short-latency
excitatory or inhibitory inputs were recorded in neurons of
the contralateral spinal dorsal horn, suggesting the exis-
Grant sponsor: the Hungarian National Research Fund; Grant number:
OTKA T025423; Grant number: T032075; Grant number: T043378; Grant
sponsor: the Hungarian Scientific Council on Health; Grant number: ETT
04-032/2000; Grant number: 623/2003 ETT.
*Correspondence to: Miklo´s Antal, Department of Anatomy, Histology
and Embryology, Faculty of Medicine, Medical and Health Science Center,
University of Debrecen, Nagyerdei krt 98, Debrecen H-4012, Hungary.
E-mail: antal@chondron.anat.dote.hu
Received 23 March 2004; Revised 24 June 2004; Accepted 10 August 2004
DOI 10.1002/cne.20356
Published online in Wiley InterScience (www.interscience.wiley.com).
THE JOURNAL OF COMPARATIVE NEUROLOGY 480:364–377 (2004)
© 2004 WILEY-LISS, INC.

tence of a segmental contralateral control over dorsal horn
cell activity (Fitzgerald, 1982, 1983). It has also been
reported that unilateral constriction of peripheral nerves
on the lower limb evokes bilateral electrophysiological and
behavioral changes (Attal et al., 1990; Colvin et al., 1996),
bilateral cell death (Sugimoto et al., 1989), and bilateral
increases in spinal cord dynorphin levels (Wagner et al.,
1993) in the lumbar spinal dorsal horn.
Providing a morphological confirmation for these func-
tional and neurochemical observations, in one of our re-
cent studies we have demonstrated that there is a sub-
stantial direct commissural connection between the
lateral aspects of laminae III–IV on the two sides of the
lumbar spinal cord (Petko´ and Antal, 2000). We have
shown that there are neurons in the lateral aspect of
laminae III–IV the axons of which cross the midline
within the posterior commissure and terminate in areas of
the contralateral dorsal horn that are identical to that in
which their cells of origin were located.
To provide further experimental evidence for a better
understanding of how these commissural neurons can be
integrated into spinal sensory circuits, in the experiments
presented here we have studied the arborization pattern,
synaptic relations, and neurochemical properties of com-
missural axon terminals.
MATERIALS AND METHODS
Animals, injection of neural tracers, and
preparation of tissue sections
Experiments were carried out on 24 adult male rats
(Wistar-Kyoto, 250–300 g, Go¨do¨llo¨, Hungary). All animal
study protocols were approved by the Animal Care and
Protection Committee at the University of Debrecen. Lam-
inectomy was performed to expose the lumbar spinal cord
under deep sodium pentobarbital anesthesia (50 mg/kg,
i.p.), while the animals were held in a stereotaxic frame.
Glass micropipettes with a tip diameter of 20–30 �m were
filled with a 10% solution of biotinylated dextran amine
(BDA; molecular weight, 10,000; Molecular Probes, Lei-
den, Netherlands) or a 2.5% solution of Phaseolus vulgaris
leucoagglutinin (PHA-L; Vector, Burlingame, CA) dis-
solved in 0.1 M phosphate buffer (PB, pH 7.4). The tracers
were injected into the lateral and medial parts of the
superficial dorsal horn at the level of the L2–L4 segments
of the lumbar spinal cord unilaterally by iontophoresis,
using a positive direct current of 5 �A with a pulse dura-
tion of 7 seconds followed by 3-second intervals for a
period of 20 minutes. Efforts were made to minimize an-
imal suffering during surgical procedures.
Following a postoperative survival period of 1 week (in
the case of BDA injections) or 3 weeks (in the case of
PHA-L injections), the animals were reanesthetized with
an overdose of sodium pentobarbital (70 mg/kg, i.p.) and
transcardially perfused first with Tyrode’s solution (oxy-
genated with a mixture of 95% O2, 5% CO2) and then with
a fixative containing 1) 4% paraformaldehyde; or 2) 4%
paraformaldehyde, 0.1% glutaraldehyde, and 0.2% picric
acid; or 3) 2.5% glutaraldehyde, 0.5% paraformaldehyde,
and 0.2% picric acid in 0.1 M PB. The lumbosacral seg-
ments of the spinal cord were removed, postfixed in the
same fixative for 1–2 hours, and immersed in 10% and
20% sucrose dissolved in 0.1 M PB until they sank. To
enhance reagent penetration, the removed spinal cord was
freeze-thawed in liquid nitrogen, sectioned at 60 �m on a
Vibratome, and extensively washed in 0.1 M PB.
Histochemical detection of the tracers
BDA. For histochemical detection of the injected BDA,
free-floating sections of the spinal cord were incubated
with avidin-biotinylated horseradish peroxidase complex
(ABC; Vector, diluted 1:100) overnight at 4°C.
PHA-L. To detect the injected PHA-L, free-floating
sections of the spinal cord were first incubated with bio-
tinylated anti-PHA-L made in goat (Vector, diluted
1:1,000) for 2 days at 4°C. The sections were then trans-
ferred into an ABC solution (diluted 1:100) for 4 hours at
room temperature.
The histochemical reactions, in both cases, were com-
pleted with a nickel-intensified diaminobenzidine (3,3�-
diaminobenzidine tetrahydrochloride; Sigma-Aldrich, St.
Louis, MO) chromogen reaction (deep blue or black reac-
tion end product; Hancock, 1984). The sections were then
extensively washed, mounted on gelatin-coated slides, and
covered with DPX neutral medium.
Preembedding nanogold
immunocytochemistry for electron
microscopy
To detect BDA-labeled and enkephalin-immunoreactive
axon terminals and perikarya in the same section, we
combined the histochemical detection of BDA with a
preembedding immunocytochemical protocol.
Following extensive washes in 0.1 M PB and treatment
with 1% sodium borohydride for 30 minutes, free-floating
sections from animals fixed with 4% paraformadehyde,
0.1% glutaraldehyde, and 0.2% picric acid were first incu-
bated with rabbit antiserum directed against methionine-
enkephalin (Peninsula, Belmont, CA, diluted 1:5,000) for
48 hours at 4°C. Then the sections were transferred into a
solution of goat anti-rabbit IgG conjugated to 1-nm gold
particles (AuroProbe One GAR, Amersham, Bucking-
hampshire, England, diluted 1:100) for 6 hours at room
temperature. After repeated washing in 0.01 M Tris-
buffered isotonic saline (TBS, pH 7.4), the sections were
postfixed for 10 minutes in 2.5% glutaraldehyde and
washed again in 0.01 M TBS and 0.1 M PB. The gold
labeling was intensified with a silver enhancement re-
agent (Aurion R-GENT, Aurion, Wageningen, Nether-
lands). To detect BDA labeling, the sections were then
transferred into an ABC solution (diluted 1:100) for 4
hours at room temperature. The ABC reaction was com-
pleted with a diaminobenzidine chromogen reaction. Sec-
tions were treated with 1% OsO4 for 45 minutes, then
dehydrated, and flat-embedded into Durcupan ACM resin
(Fluka Chemie, Buchs, Switzerland) on glass slides. Se-
lected areas were reembedded, and serial ultrathin sec-
tions were cut and collected on Formvar-coated single-slot
nickel grids.
Double and triple immunostaining for
confocal microscopy
To investigate whether BDA-labeled commissural axon
terminals express neural markers like vesicular gluta-
mate transporters (VGLUTs), glutamic acid decarboxyl-
ase (GAD), or glycine transporter (GLYT), which are char-
acteristic of various subsets of excitatory and inhibitory
neurons, the immunoreactivity of VGLUT1, VGLUT2,
365COMMISSURAL PROPRIOSPINAL PATHWAY IN THE DORSAL HORN

VGLUT3, GAD65/67, and GLYT2 of BDA-labeled axon
terminals was tested on free-floating sections from ani-
mals fixed with 4% paraformadehyde.
Double immunostaining. Sections were first incu-
bated with one of the following antibodies: antibody
against VGLUT1 raised in guinea pig (diluted 1:5,000;
Chemicon, Temecula, CA), antibody against VGLUT2
raised in guinea pig (diluted 1:2,000; Chemicon), antibody
against VGLUT3 raised in guinea pig (diluted 1:5,000;
Chemicon) for 2 days and then transferred into a solution
that contained goat anti-guinea pig IgG conjugated with
Alexa Fluor 488 (diluted 1:500) for 5–6 hours. The sec-
tions were finally treated with streptavidin conjugated
with Alexa Fluor 546 (diluted 1:2,000; Molecular Probes)
to detect BDA.
Triple immunostaining. Sections were first incu-
bated with a mixture of antibodies against GAD65/67
raised in rabbit (diluted 1:2,000; Sigma, St. Louis, MO)
and antibody against GLYT2 raised in sheep (diluted
1:4,000; Chemicon) for 2 days. After extensive washing,
the sections were incubated for 5–6 hours in donkey anti-
rabbit IgG conjugated with Alexa Fluor 647 (diluted 1:500;
Molecular Probes) and donkey anti-sheep IgG conjugated
with Alexa Fluor 555 (diluted 1:500; Molecular Probes)
and were then transferred into a solution that contained
streptavidin conjugated with Alexa Fluor 488 (diluted
1:2,000; Molecular Probes) for 1 hour to detect BDA.
Before the antibody treatments, the sections were kept
in 20% normal goat serum (Vector) for 50 minutes. Anti-
bodies were diluted in 10 mM TBS to which 1% normal
goat serum (Vector) was added. Sections were mounted on
glass slides and covered with Vectashield (Vector).
As controls for the specificity of the immunostaining
procedure, some sections were incubated in normal goat
serum (1:100) instead of the primary antisera. No staining
was observed in these sections.
Postembedding immunocytochemistry for
GABA and glycine
Alternate Vibratome sections processed for histochemi-
cal detection of PHA-L were treated with 1% OsO4 for 45
minutes, dehydrated, and flat-embedded into Durcupan
ACM resin (Fluka Chemie) on glass slides. Selected areas
containing labeled terminals were reembedded, and serial
ultrathin sections were cut and collected on Formvar-
coated single-slot nickel grids. Sections on consecutive
grids were processed for �-aminobutyric acid (GABA) and
glycine immunocytochemistry, whereas sections on every
third grid were counterstained with lead citrate. The im-
munostaining was performed according to the postembed-
ding immunogold procedure described by Somogyi and
Hodgson (1985). Sections were treated with 1% periodic
acid for 10 minutes and then transferred onto 2% sodium
periodate for 10 minutes. Following treatment with 1%
ovalbumin for 30 minutes, sections were incubated with
rabbit anti-GABA (code: 416; diluted 1:1,000) or rabbit
anti-glycine (code: 290; diluted 1:800) for 90 minutes.
The immunocytochemical characteristics of the antisera
have been extensively tested and published earlier (Kol-
ston et al., 1992). The primary antisera were generated to
the amino acid coupled to protein by glutaraldehyde ac-
cording to the principles devised by Storm-Mathisen et al.
(1983). To abolish possible weak cross-reactivities with
other amino acids, the anti-GABA antiserum was prein-
cubated in 400 �M glutamate-glutaraldehyde conjugates,
and the anti-glycine antiserum was preincubated in 100
�M GABA- and glutamate-glutaraldehyde conjugates,
prepared as described by Dale et al. (1986). Subsequently
the grids were transferred onto goat anti-rabbit IgG-
coated colloidal gold (15 nm, BioCell, Cardiff, UK; diluted
1:20) for 2 hours. Sections were counterstained with ura-
nyl acetate and lead citrate.
Antiserum specificity was tested by treating the diluted
anti-GABA and anti-glycine sera with glutaraldehyde-
GABA and glutaraldehyde-glycine conjugates, respec-
tively. The conjugation of GABA and glycine with glutar-
aldehyde was performed according to the method of Dale
et al. (1986). Ultrathin sections were incubated according
to the immunocytochemical procedure by using the
glutaraldehyde-GABA and glutaraldehyde-glycine conju-
gates as primary sera. Under these conditions specific
immunostaining was completely abolished.
Immunoreactivity for GABA and glycine was evaluated
by a quantitative criterion. The sectioned area of PHA-L–
labeled terminals and their postsynaptic structures were
measured from electron micrographs of GABA- or glycine-
reacted sections by a planimeter, and the unit density of
immunogold particles was calculated. The density of im-
munogold particles was also measured over synaptic bou-
tons forming asymmetric synapses within the same elec-
tron micrograph. Terminals forming asymmetric synaptic
contacts are not thought to contain GABA in the spinal
cord; therefore the average value calculated per unit area
on each micrograph was considered to represent nonspe-
cific background staining in the procedure. Labeled termi-
nals and their postsynaptic structures with gold particle
density at least five times and three times the background
were considered to be GABA- and glycine-immunorective,
respectively.
Photomicrograph production
Sections stained for conventional light microscopy were
observed in a Nikon Eclipse 800 light microscope. Digi-
tized images were captured by using a SPOT RT Slider
Fig. 1. a: Photomicrograph showing an injection site of PHA-L in
the lateral part of the dorsal horn. b: Camera lucida drawing of a
transverse section of the lumbar spinal cord showing the sites and
sizes of PHA-L (bordered by continuous lines) and BDA (bordered by
dashed lines) injections into the lateral and medial aspects of the
dorsal horn. The borders of the gray and white matters as well as the
cytoarchitectonic laminae of the gray matter are drawn by dashed and
continuous lines, respectively. Roman numerals indicate the Rexed
laminae of the gray matter. c: Camera lucida drawing and photomi-
crograph illustrating the commissural connection between the lateral
aspects of the dorsal horns on the two sides of the spinal cord. The
hatched area on the camera lucida drawing and the darkly stained
territory on the micrograph label an injection site of BDA. After
arising from the injection site, a fiber bundle crosses the midline
dorsal to the central canal, turns into the dorsolateral direction,and
arborizes in the lateral aspect of laminae III–IV on the opposite side
of the spinal cord. d,e: Photomicrographs showing the cell bodies and
dendritic arbors of commissural neurons in horizontal sections of the
lateral aspect of laminae III–IV. The neurons were retrogradely la-
beled with BDA. f: Composite camera lucida drawing of the trans-
verse section of the lumbar spinal gray matter illustrating distribu-
tion of commissural neurons retrogradely labeled with BDA in the
dorsal horn. Camera lucida drawings of 10 individual sections are
superimposed. Note that most of the labeled neurons are located in
the very lateral aspect of laminae III–IV. Scale bar � 100 �m in a,c;
50 �m in d,e.
366 M. PETKO´ ET AL.

Figure 1

camera. The immunofluorescence-labeled sections were
examined in a Zeiss LSM 510 laser scanning confocal
microscope, and digital images were captured by using the
Zeiss LSM software. Sections for electron microscopy were
investigated with a JEOL 1010 TEM. Images were re-
corded on photographic negative slides that were then
digitized with a Nikon Super Coolscan 8000 scanner by
using the Nikon Scan 3.1 software.
In case of both light and electron microscopy, digitized
images were stored in an IBM PC. Photoshop 7.0 was used
for final adjustments of contrast and brightness in all
pictures.
RESULTS
Injection sites
The tracers were delivered into the lumbar spinal cord
at the level of L2–-L4 segments. The iontophoretic injec-
tions of both BDA and PHA-L yielded small and well-
defined injection sites that involved ovoid cross-sectional
areas with a mediolateral extent of 100–300 �m and ex-
tended across laminae I–III, with some involvement of
lamina IV (Fig. 1a–c). As detected by the histochemical
procedure, many cells incorporated the tracer within the
confines of the areas infiltrated by the tracers. The ap-
pearance of these cells, as well as that of the injection
sites, was similar to those reported in previous studies
(Gerfen and Sawchenko, 1984; Wouterlood and Groenewe-
gen, 1985; Veenman et al., 1992; Rajakumar et al., 1993).
The experiments were carried out on 24 animals. Be-
cause of inappropriate localization of the injection sites,
seven animals had to be excluded from the final evalua-
tion. Consequently, data presented in this study are based
on the results obtained in 17 animals. On the basis of the
applied tracer and the location of the injection site, the
animals were divided into the following four experimental
groups:
1. PHA-L injection into the lateral regions of laminae
I–IV (three animals, Fig. 1a,b)
2. BDA injection into the lateral part of laminae I–IV
(eight animals, Fig. 1b,c)
3. PHA-L injection into the medial areas of laminae I–IV
(three animals, Fig. 1b)
4. BDA injection into the medial aspect of laminae I–IV
(three animals, Fig. 1b).
Distribution and morphology of commissural
neurons and their axon terminals
Injections of PHA-L and BDA into the lateral areas of
laminae I–IV resulted in extensive labeling of axons and
axon terminals in the lateral aspect of the contralateral
dorsal horn, mostly in laminae III–IV (Fig. 1c). The cross-
ing fibers arose from the lateral dorsal horn (Fig. 1c). Most
of them ascended or descended for various distances (from
a few hundred micrometers to four spinal segments) from
the injection site; then the fibers turned medially and
crossed the midline in small bundles, forming a ladder-
like pattern along the rostrocaudal axis of the spinal cord.
On the contralateral side they reached the lateral aspect
of the dorsal horn and terminated there (Fig. 1c), forming
a dense termination field mostly in laminae III–IV.
When BDA was injected into the lateral dorsal horn, in
addition to the axon terminals, retrogradely labeled neu-
rons were also recovered in the lateral aspects of the
contralateral dorsal horn. Most of the retrogradely labeled
neurons were confined to the very lateral aspect of lami-
nae III–IV (Fig. 1f), where the terminals of the antero-
gradely stained axons were found. Stained cells were re-
covered at other locations, e.g., in the lateral part of
lamina II or in the medial divisions of laminae II–IV, in
very limited numbers (Fig. 1f). Most of the labeled neu-
rons presented somatodendritic morphology that resem-
bled the morphology of either islet or stalked cells in the
superficial spinal dorsal horn (Fig. 1d,e). Some of them
had spheroid (Fig. 1d) and others fusiform cell bodies (Fig.
1e) with a diameter of 10–25 �m. Two main dendrites
arose from the cell bodies either from the two poles of the
fusiform perikarya or from the lateral aspect of the spher-
oid somata. The dendritic arbors extended along the ros-
trocaudal axis of the spinal cord, whereas the mediolateral
and dorsoventral extensions of the dendritic tree were
restricted.
After tracers were injected into the medial aspects of the
dorsal horn, commissural fibers were only occasionally
observed. These axons crossed the midline in the posterior
commissure and terminated in the medial half of the dor-
sal horn.
Postsynaptic targets of commissural axon
terminals
Ninety-one PHA-L–labeled terminals with highly iden-
tifiable synaptic contacts were investigated at the ultra-
structural level. The outline of the profiles was dome
shaped or slightly oval, with a diameter of 0.3–1.5 �m
(Fig. 2). Of the 91 stained axon terminals, 67 (73.6%)
formed symmetric (Fig. 2a,b,d) and 24 (26.4%) asymmetric
(Fig. 2c,e) synaptic appositions. Most of the labeled termi-
nals were engaged in synaptic contacts with dendrites
(Fig. 2a,b,e), the diameter of which varied in the range of
0.4–3.0 �m. Only eight (8.8%) and four (4.4%) of them
established axospinous (Fig. 2c) and axosomatic (Fig. 2d)
synaptic contacts, respectively.
In the case of BDA injections, anterogradely labeled
axon terminals and retrogradely labeled neurons ap-
peared simultaneously in the investigated sections, which
allowed us to look for interactions between commissural
neurons. By careful examination of the material, we found
that a number of stained axon terminals established syn-
aptic contacts with dendrites of labeled neurons (Fig. 2e).
Met-enkephalin, VGLUT1, VGLUT2, VGLUT3,
GAD65/67, GLYT2, GABA, and glycine
immunoreactivity of commissural axon
terminals
The preembedding nanogold technique labeled a large
number of axon terminals for met-enkephalin (Fig. 3a).
Despite the strong immunostaining of the sections, none
of the BDA-labeled axon terminals had a positive immu-
noreaction for met-enkephalin (Fig. 3b).
Confirming results of previous studies (Li et al., 2003;
Oliveira et al., 2003), axon terminals immunoreactive for
VGLUT1, -2, and -3 were revealed in various numbers and
distributions in the dorsal horn. In the lateral aspect of
laminae III–IV, where most of the BDA-labeled commissural
propriospinal terminals were distributed, axon terminals
immunoreactive for VGLUT1, -2, and -3 were all scattered,
but all three forms of VGLUTs showed a complete segrega-
tion from BDA-labeled axon terminals (Fig. 4).
368 M. PETKO´ ET AL.