Androgen alters the dendritic arbors of SNB motoneurons by acting upon their target muscles.

Abstract

In adult male rats, motoneurons of the spinal nucleus of the bulbocavernosus (SNB) have been shown to retract and reextend their dendritic branches in response to systemic androgen deprivation and readministration. Furthermore, other studies have suggested that the dendritic complexity of neurons can be regulated by their targets. To assess whether androgens might act upon the target muscles to mediate changes in SNB dendrites, adult male rats were castrated and implanted with a small capsule filled with testosterone (T) next to the bulbocavernosus and levator ani muscle complex (BC/LA) on one side, while the muscles on the contralateral side were implanted with another capsule containing hydroxyflutamide (hFl), an anti-androgen. We have previously shown that after 30 d of this focused, lateralized androgen treatment the BC/LA complex is significantly larger on the T-treated side. We now report that the total dendritic lengths of SNB motoneurons innervating muscles given androgen blockade are reduced by 44% compared to SNB motoneurons innervating muscles given androgen stimulation. Dendrite lengths within three regions of the spinal cord were altered in a nonuniform manner: large changes occurred in the dorsal and contralateral dendritic fields while there was no difference in the ipsilateral dendritic field. These results suggest that BC/LA muscles, in response to androgen stimulation, produce a trophic substance which regulates the dendritic organization of SNB motoneurons in adulthood.

Full text

Page 1

The Journal of Neuroscience, June 1995, 15(6): 4406-4416
Androgen
Acting
Alters the Dendritic
upon Their Target Muscles
Arbors of SNB Motoneurons by
Mark N. Rand” and S. Marc Breedlove
Department of Psychology and Graduate Groups in Neurobiology and Endocrinology, University of California,
Berkeley, California 94720-l 650.
In adult
the bulbocavernosus
and reextend
temic androgen
more, other studies
plexity of neurons
sess whether
to mediate
castrated
testosterone
ani muscle
cles on the contralateral
capsule containing
gen. We have previously
cused, lateralized
is significantly
that the total dendritic
vating muscles
44% compared
given androgen
regions of the spinal
manner: large
lateral dendritic
ipsilateral dendritic
muscles, in response
trophic substance
tion of SNB motoneurons
[Key words:
sus, levator
male rats, motoneurons of the spinal
been
branches
and readministration.
have suggested
can be regulated
androgens might act upon the target
changes in SNB dendrites,
and implanted with
(T) next to the bulbocavernosus
complex (BC/LA)
side were
hydroxyflutamide
shown
androgen treatment
larger on the T-treated
lengths
given androgen
to SNB motoneurons
stimulation. Dendrite
cord were
changes occurred
fields while there was no difference
field. These
to androgen
which regulates
in adulthood.
neural plasticity,
ani, testosterone,
nucleus
to retract
to sys-
Further-
of
(SNB) have shown
in response their dendritic
deprivation
that the dendritic
by their targets.
com-
To as-
muscles
adult male rats were
capsule
and levator
on one side, while
implanted
(hFI), an anti-andro-
that after 30 d of this fo-
the BCYLA complex
side. We now report
of SNB motoneurons
blockade are reduced
innervating
lengths
altered in a nonuniform
in the dorsal
a small filled with
the mus-
with another
inner-
by
muscles
three within
and contra-
in the
results suggest that BCYLA
produce
organiza-
stimulation,
the dendritic
a
dendrites,
hydroxyflutamide]
bulbocaverno-
The perineal muscle complex of the bulbocavernosus
ani (BC/LA) is sexually
but greatly reduced in female rats (Hayes,
1970; Joubert and Tobin,
rons from the spinal nucleus
(Breedlove and Arnold,
rons innervating the BC/LA
development by exposure
and levator
dimorphic, being present in male rats
1965; Cihak et al.,
1989), and is innervated
of the bulbocavernosus
1980; Schreder, 1980). SNB motoneu-
are spared from cell death during
to androgens (Breedlove
by motoneu-
(SNB)
and Arnold,
Received
We thank
for comments
NS2X321.
Correspondence
Psychology.
1650.
‘Present
biology.
Copyright
Oct. 3. lYY4;
Lnurcn
on
revised Jun. 13,
the dendritic
This
1995: accepted Jan.
and
19,
Dr. Nancy
by
1995.
Levitt
the
fur tracing
manuscript.
arbors Forger
Grant work was supported NIH
should
Tolman
be addressed
Hall,
to S. Marc
of California,
Breedlove, Department
CA
of
3210 University Berkeley, 94720.
address:
33.1 Cedar
0
Yale
Street,
Society
University
West
for Neuroscience
School of Medicine, Section of Neuro-
Haven, CT 06510.
I995 0270.6474/95/154408-09$05.00/O
1983; Nordeen
targets. In adult males, reduced
panied by decreases in SNB soma size (Breedlove
198 I), size and number of synapses upon SNB somas and prox-
imal dendrites (Leedy et al., 1987), dendritic
1986; Forger and Breedlove, 1987; Sasaki and Arnold,
and BC/LA muscle mass (Venable,
are present in SNB motoneurons
BC/LA muscles (Jung and Baulieu,
ferents (Monaghan and Breedlove,
of these tissues androgens might act upon to induce changes in
SNB motoneuron morphology.
We have previously reported
fect upon BCYLA muscle size (Rand and Breedlove,
that study adult male rats were gonadectomized
treatment was provided to one side of the BC/LA
the other side was treated with hydroxyflutamide
anti-androgen. After 30 d the BC/LA
treated side was consistently heavier and had larger muscle fiber
diameters than its contiguous, contralateral
ing that the anabolic effect of androgen
the muscle. This preparation also offered an opportunity
vestigate potential changes in motoneuron
from androgen treatment of the target muscles; SNB motoneu-
rons on one side of the spinal cord innervate
lated targets, while the contralateral
pm away innervate androgen-deprived
hormone from the treatment site at the BC/LA
bilateral effects on distant tissues-in
spinal cord. Using this experimental
that androgenic stimulation of BC/LA
motoneuron dendrites.
et al., 1985), which also masculinize
systemic androgens
the muscle
are accom-
and Arnold,
extent (Kurz et al.,
1991),
1966). Androgen
(Breedlove
1973), and supraspinal
1991). It is not known
receptors
I980), and Arnold,
af-
which
that androgens have a direct ef-
1992). In
and androgen
complex while
(hFl), a potent
on the androgen- complex
counterpart,
is mediated
indicat-
at or near
to in-
resulting morphology
androgen-stimu-
SNB motoneurons
targets. Any diffusion
would have equal
particular
technique, we now report
muscles can alter SNB
only 300
of
the brain and
Materials and Methods
Cupde
viously described (Rand and Breedlove, 1992). Briefly, a reservoir was
made from a short piece of Tygon tubing which was heat-sealed at one
end and filled with either crystalline testosterone (T; Sigma Chemicals,
St. Louis) or the anti-androgen hydroxyflutamide
courtesy of Dr. R. Neri, Schering Corp.). The open end of the reservoir
was then sealed with liquid SILASTIC
hormone diffusion, apprbximately 2 mm’. The reservoir was placed in
a mold and embedded in liauid SILASTIC.
and formed flaps for suturihg the capsule to the muscle (Fig. I).
Adult male Sprague-Dawley rats 90-120 d of age were deeply anes-
thetized and implanted with pairs of capsules, one on each side of the
BC/LA complex. Animals were castrated at the time of capsule im-
plantation; a T capsule was sutured onto one side of the BC/LA com-
plex, selected at random, and an hF1 capsule was sutured onto the con-
tralateral BC/LA, as far apart as possible (15-20 mm). At this time,
animals were ear-marked and their records were coded so that all sub-
preparutimz
and inzplanrution. Capsules were prepared as pre-
(hFI; SCH16423,
to provide a uniform area for
which encased the reservoir

Page 2

The Journal of Neuroscience, June 1995, 7~76) 4409
Figure
to androgen-sensitive
B, Bottom view. This surface was next to the muscle; hormone diffusion
occurred from the circular area at the center (urrow). The capsules were
sutured onto the bulbocavernosus/levator
in bilateral pairs; one contained testosterone while the capsule on the
contralateral side contained hydroxyflutamide,
resulted in strong local androgen stimulation to one side of the BULA
muscle complex and relatively little androgen stimulation to distal tis-
sues such as the brain and spinal cord.
1. Capsule used to deliver a local gradient of steroid treatment
muscles; overall length is 10 mm. A, Top view.
ani (BULA) muscle complex
an anti-androgen. This
sequent work was done without knowledge of the T and hFl treatment
sides. Thirty days after gonadectomy and capsule implantation
imals were anesthetized and given a single I pl injection of 0.2% chol-
era toxin-horseradish peroxidase conjugate (CT-HRP; List Laboratories,
Cupertino, CA) into the BC on one side. The side of CT-HRP injection
was chosen randomly, and the injection was made into the muscle be-
neath the capsule. Capsules were verified for lateral placement and
structural integrity, and then removed. Animals were sacrificed 90-100
hr after CT-HRP injection.
Histological procedures.
Animals were anesthetized with an overdose
of pentobarbital and perfused intracardially
lowed by 200 ml of buffered 1% paraformaldehyde
aldehyde, then 500 ml of 10% phosphate-buffered sucrose. Spinal cords
were removed and stored at 4°C overnight in 10% phosphate-buffered
sucrose. The BC/LA muscles and seminal vesicles were removed and
stored in buffered 4% paraformaldehyde
the muscles were trimmed of fat, connective tissue, ischium and ischio-
cavernosus remnants, bulbourethral
(Fig. 2A). Left and right BC/LA muscle complexes were divided with-
out knowledge of their differential hormone treatment; the medial BCs
were separated at their point of attachment to the base of the penis, the
base of the penis was cut in half longitudinally,
divided at the median raphe (Fig. 2B). BC/LA halves and seminal ves-
icles were weighed to the nearest milligram.
experimental animals, two other groups of adult males of equivalent
age and body weight were sacrificed, perfused, and their BULA
cles and seminal vesicles were removed and weighed. One group had
been castrated 30 d earlier without having capsules implanted (n = 7);
the other group was gonadally intact and untreated (n = 18).
The day after sacrifice spinal cords were frozen sectioned in the trans-
verse plane at a thickness of 50 pm. Every section was taken over a 6
the an-
with 50 ml of saline fol-
+ 1.25% glutar-
for 2 weeks. After this time
glands, rectum, and the distal penis
and the LAS were
For comparison with the
mus-
mm length (i.e., 2 mm beyond the rostra1 and caudal borders of the
SNB). Sections were reacted with tetramethybenzidine
scribed by Mesulam (1978) to visualize the CT-HRP Spinal cords were
processed in batches of three or six; each batch contained a random
mix of animals with respect to androgen treatment of the CT-HRP la-
beled side.
CT-HRP labeling of the motoneuron dendrites was rated satisfactory
or unsatisfactory in each spinal cord section without knowledge of treat-
ment at the muscle. In total, 57 animals were gonadectomized, implant-
ed with capsules and unilaterally injected with CT-HRP; of these, I9
animals met the requirements for further analysis, that is, the capsules
remained laterally placed during the treatment interval and were struc-
turally intact, and the spinal cords were well perfused and had clearly
labeled motoneurons with little background reaction product. Twelve of
these animals had received CT-HRP injections in the T-treated muscles,
and seven animals had been injected on the hFl-treated side.
Measurement and analysis of dendrites.
the dendritic arbors photomicrographs
sections, dendrites were traced from the prints, and the tracings were
scanned into a microcomputer. This resulted in a relatively permanent
record of the arbors, which was important because the TMB reaction
product can degrade over time, and also allowed counterstaining the
tissue (which tended to wash out the reaction product) without
nently losing information. This method also provided a complete record
at each stage of measurement and greatly facilitated review of the ac-
curacy of traced processes before digitization.
low numeric aperture plan
objective almost all of the dendrites within
the 50 ym thick sections were in focus (see Fig. 5). The data obtained
with this method are comparable to other studies of SNB dendrites using
CT-HRP retrograde Iabelling (Kurz et al., 1986; Goldstein et al., 1990;
after split cell body correction for both studies), but are not directly
comparable to three-dimensional reconstruction studies (Sasaki and Ar-
nold, 1991) due to loss of information in the z-plane. Like other CT-
HRP studies (e.g., Kurz et al., 1986; Goldstein et al., 1990), our two-
dimensional measures of dendritic arbor length were used to evaluate
relative changes in dendrites within the transverse plane and do not
represent absolute arbor lengths.
Spinal cord sections were examined with a microscope under dark-
field illumination and each section which contained labeled dendrites
was noted (typically 3&40 sections per animal). Later, 15 of the labeled
sections were randomly selected, relocated and photographed using a
Zeiss 6.3X PlanApo Neofluar 0.20 N.A. objective and a IOX monoc-
ular. The number of labelled cell bodies in each photographed section
was counted under the microscope and recorded. Prints were made at
a final magnification of 50X, in batches of six animals at a time. Ten
of the 15 prints from each animal were then selected
clarity of the photograph and were used to trace the dendritic processes
and determine the number of labeled cell bodies.
SNB dendrites form distinct projection fields in the ipsilateral, con-
tralateral, and dorsal areas of the transverse plane of the spinal cord. To
divide the transverse plane, a template with an origin point and three
lines was made on a clear sheet of acetate and placed over each pho-
tomicrograph (Fig. 3). The origin point was initially
“best fit” to the center of the SNB nucleus in all the photographs, and
was set 280 pm from the central canal and 140 pm from the midline.
The dorsal field spanned 150”, whereas the ipsilateral and contralateral
fields were 105” each. The dendrites within each field were drawn on
another clear acetate sheet which was placed over the template and
photograph. All drawings were made by an individual who was not
aware of the nature of the experimental treatments.
Dendrites were traced using a 0. I3 mm tip drafting pen to retain as
much detail as possible and then scanned into a microcomputer as bi-
nary image files. The digitized images were analyzed using NIH Image
(Wayne Rasband, National Institutes of Health; NIH Image is available
via anonymous FTP from zippy.nimh.nih.gov).
due to pen line width each image was “skeletonized,”
age processing function which removes pixels from lines until they are
only 1 pixel wide. This procedure also eliminated any extra pixels at
the junction of crossed dendrites, which would have otherwise resulted
in somewhat overestimated measures of arbors whose dendrites crossed
with greater frequency. The number of black pixels in each image was
counted by the program and converted into a measure of linear distance.
The conversion factor was derived from a skeletonized scanned image,
drawn with the same pen, consisting of 16 lines of equal length regu-
larly
(TMB) as de-
To measure the lengths of
were made of the spinal cord
perma-
By using a low-power,
based upon the
determined by
To reduce variability
an iterative im-
spaced in a radial array.

Page 3

4410 Rand and Breedlove * Motoneuron Morphology and Androgen Stimulation of Target Muscles
Figure 2. Bulbocavernosus (BC) and levator ani (LA) muscle complex after 30 d of lateralized androgen treatment. The BC/LA muscles on the
left side of the photographs were given testosterone treatment, while the BC/LA muscles on the right side were given hydroxyflutamide
androgen) treatment; the muscle complex is visibly larger on the testosterone-treated side. A, Ventral view. B, Dorsal view. There is a prominent
difference in the size of the LA muscles on either side of their junction, the median raphe (arrow). Scale bar, IO mm.
(anti-
The ipsilateral field contains the axons of the filled SNB motoneu-
rons, which are often located within a bundle of dendrites. Post hoc
examination of the photomicrographs
the density of dendritic bundling or in the number of dendritic bundles
by treatment. To accurately measure changes in dendrite lengths in the
ipsilateral field the length of the filled axons was estimated and sub-
tracted. Axons typically make a characteristic “hook”
trolateral spinal cord (Sasaki and Arnold,
counting the number of hooks in each section we estimated the number
of labeled axons. This number correlated I : 1 with the corrected number
of cell bodies 0, = 0.97x + 0.65; r> = 0.70; p < 0.0001). indicating
showed no obvious differences in
out of the ven-
1991; also see Fig. 5); by
Figure 3.
cord. The traced dendrites of retrogradely labeled SNB motoneurons
within a sampled section are shown. In this plane the dendrites typically
distribute within three fields: A, contralateral; B, dorsal; and C, ipsilat-
era1 to the SNB cell bodies. A template was made to divide the trans-
verse plane into these areas, with an origin point set 140 brn from the
midline and 280 pm from the central canal.
Diagram of a transverse section through L5 in the rat spinal
that most if not all of the SNB motoneurons possessed hooked axons.
Six &xons which could be clearly distinguished
which were contiguous within the 50 pm thick sections were drawn
and measured, and the average was used to estimate a standard length
for subtraction. Axons accounted for approximately
processes in the ipsilateral field and did not affect the statistical outcome
when subtracted. It was not possible to identify and subtract recurrent
axon collaterals, which occur independently of the hooked axonal sig-
natures and are seen in approximately
(Sasaki and Arnold, 1991).
In addition to calculating the total length of dendrites in each field,
the maximum distance of label within dendrites (measured linearly from
the cell body origin point) was determined for each field in each spinal
cord section, and averaged for each animal.
h4easurement
of SNB sumu areas. After photographing
processes, the somas of the SNB motoneurons were stained and mea-
sured. Coverslips were removed with xylene and the spinal cord sec-
tions were cleared of mounting media and stained with thionin. Thionin-
stained SNB cells were identified by their characteristically large, mul-
tipolar cell bodies located in the dorsal area of the ventromedial spinal
cord. Cell bodies were traced using a Zeiss Standard microscope
equipped with a drawing tube at a final magnification of 650X. Cells
were traced only if they had a clearly visible nucleus, and if there was
sufficient CT-HRP label in the SNB cell bodies to identify the side of
the injection (dendrites extend contralaterally to the opposite SNB nu-
cleus; without labeled somas the source of the dendrites could be am-
biguous). Fifteen thionin-stained somas on each side of the spinal cords
of 16 animals were traced. The drawings were scanned into a computer
and their areas were measured with NIH IMAGE.
Corrected cell counts and dendritic length determinations.
SNB cells were corrected for split cell number in each animal by using
the average CT-HRP-filled soma size of the animal in a count correction
formula (method of Konigsmark, 1970). For each subject the dendritic
arbor lengths for each field were summed, these field totals were
summed, and then all totals were divided by the corrected number of
cell bodies to obtain an estimate of total dendritic arbor length per cell
and average length of dendrites per field per cell.
Statistical analyses. Statistical analyses were made for the effects of
treatment upon seminal vesicle weights, muscle complex weights, total
dendritic arbor lengths, SNB soma size, maximum dendrite lengths, and
differences in the amount of dendrites within the ipsilateral, contralat-
from dendrites and
19% of the filled
25% of all SNB motoneurons
the dendritic
Labeled

Page 4

The Journal of Neuroscience, June 1995, 136) 4411
eral,
described
centage
T treatment
normally-present
relative
and dorsal
in
differences
fields of the spinal
errors
cord.
of
the
considered
differences
Tests
means
equation
and comparisons
given.
(T -
attempt
therefore
are
per- Results; standard
were
gonadectomy
androgens,
T-treated
the are All
derived from
was
/IF/)/~.
to
Local
after an maintain
calculated and were
to the side.
As expected, there
by CT-HRP
Animals
side had seminal
were no significant
labeled side (between-subjects
CT-HRP labeled
vesicles which
differences in seminal
vesicle
two-tailed).
T-treated
mg (w = I2), while animals with CT-HRP
on the hFI-treated side had seminal vesicles which
? I2 mg (II = 7). These data indicate
levels were uniform across the experimental
Overall, seminal vesicles
weighed I64 + I7 mg (n = l9), not significantly
a group of males castrated
II = 7; between-subjects
a group of gonadally intact males in the same body weight
(n = IX) had an average seminal
mg. approximately seven times heavier than the untreated
trated animals or the experimental
jects
t test, two-tailed; 17 < 0.0001
greatly reduced seminal vesicle weights
imals demonstrated that systemic androgen
and that little if any T escaped from the capsules into systemic
circulation. Prior experiments
planted with blank capsules and T capsules,
and hFI capsules also found that capsule treatment
no systemic effects (Rand and Breedlove,
weights
t test,
on with motoneurons
weighed
labelled
the
166 + 26
motoneurons
weighed 15X
that systemic androgen
group.
the experimental from animals
from
19
different
(I 54
In contrast,
30 d before sacrifice + mg,
t test, two-tailed; Fig. 4A).
range
+ 35
cas-
vesicle weight of II64
T/hFI
in both
group
cases;
(between-sub-
Fig.
4A).
The
an- of the experimental
levels were quite low
using animals which
or blank capsules
had virtually
1992).
were im-
BC/LA
heavier
blockade
Fig. 4B),
1992). The average difference
was 17% in the present study. T-treated
? 30 mg, and hFl-treated
These values were intermediate
animals and animals
sacrifice (Fig. 4B).
muscle halves given local T treatment
than the contralateral
(within-subjects
t test, two-tailed,
replicating previous
were significantly
given local
n = 19; 17 < 0.0001;
(Rand and Breedlove,
between the two sides
BClLAs weighed
weighed 457
to the BC/LA weights of normal
were gonadectomized
BC/LAs androgen
results
in weight
542
BC/LAs + 36 mg.
which 30 d prior to
Soma areas did not differ by treatment at the muscles (Fig. 4C),
by comparison of either CT-HRP-tilled
subjects t test, two-tailed) or thionin-stained
in-subjects
t test, two-tailed). As might
filled more of the cell body than did the Nissl stain. A power
analysis indicated that the sample sizes were more than adequate
to detect even half of the effect of systemic androgens
size reported previously (Breedlove
number of CT-HRP labeled motoneurons
cle treatment (T-side = 12.3 t 0.9, hFI-side
motoneurons
motoneurons
be expected,
(between-
(with-
CT-HRP
on soma
I98 I ). The
by
and Arnold,
did not
= I I. I ? I .7).
differ mus-
Totul
Figure
SNB
plane of the spinal cord; the motoneurons
vated hFl-treated BC/LA
dendritic
5 shows dark-field
motoneuron
lengths
photomicrographs of CT-HRP-filled
within the transverse
in Figure
somas and dendrites
5A inner-
in Figure muscles and the motoneurons
A
1200
loo0
&300
E
.-
mm
p 400
200
0
Seminal Vesicles
1200
1000
600
600
400
200
0
BC/LA Muscle Halves
n L
SNB Somata
1200
h loo0
@ii
s 600
8600
4
400
200
A
”
I I
CT-HRP
Thionin
Fi~llr-e
ment
parisons.
experimental
tomized
had
to gonadally
testosterone
BULA
4. Measures
BC/LA
A. Seminal
group
for the
vesicle
of systemic and local effects of androgen
statistical
treat-
con- at the muscle
vesicle
and
length
weights
normal
were
(from
complex.
weights
a group
of
which
animals.
significantly
the same
muscles
30 d GX
areas
muscles,
bodies,
H~~r&t.r
were
animals
(30
were
R,
heavier
animal)
were
group
were not
by either
or a within-subjects
indicate
different
which
rl GX).
substantially
BULA
compared
treated
intermediate
and the group
significantly
a between-subjects
not between
been
of these
reduced
muscles
to
with
in
of gonadally
different
the
of had gonadec-
proups
compared
treated
contralatcral
anti-androgen
weight
same time Both
seminal
intact,
(r)
with
muscles
The
muscles
animals.
gen treatment
of CT-HRP-filled
in-stained
(/IF/).
LA
capsule-treated
from
SNB
at the
to BC/
intact
andro-
the
soma C. by
comparison
of thion- cell comparison
cell bodies.
5B innervated
rons innervating
overall
muscles given androgen
tailed; 17 < 0.05; Fig. 6A).
ed side averaged
on the hFl-treated
These differences
arbor extents within
3), as well as the lengths of the most distally labeled processes.
T-treated
muscles given androgen
by 44% relative
BC/LA muscles. Dendrites of motoneu-
were reduced
innervating
t test, two-
on the T-treat-
body, whereas those
? Xl9 pm/cell
by comparing
blockade
to those of motoneurons
treatment (between-subjects
Arbors of motoneurons
I I, I34 & 1493 Fm/cell
side averaged
were further examined
three areas of the transverse plane (see Fig.
6279 body.
dendritic
Ipsilaterul,
A two-way
drites per field by treatment
main effect of treatment
icant main effect of field [F(2, 17) = 66.43; 17 < O.OOOl~, and
a significant interaction of treatment by field [F(2. 34) = 12.22;
p < O.OOOl]. Consequently, the amount of dendritic
each field was compared by treatment
dorsal,
repeated-measures
und contralaterul ,field dmdritic
of the amount of den-
a significant
Icwgths
ANOVA
at the muscle showed
[F( I, 17) = 5.50; /I < 0.051. a signif-
arbor within
at the muscle (between-

Page 5

4412 Rand and Breedlove
l Motoneuron Morphology and Androgen Stimulation of Target Muscles
Figure 5. Dark-field photomicrographs
spinal cord. SNB motoneurons which innervated androgen-depleted
of cholera toxin-horseradish peroxidase (CT-HRP)-filled
muscles (A) had substantially reduced dendritic arbors compared to motoneurons
SNB motoneurons in the transverse plane of the

Page 6

The Journal of Neuroscience, June 1995, 15(6) 4413
-i? A
3
6 12000
2
i 10000
0
\ 6000
b
9 6000
0
.=
‘C
4000
z
2
2OQO
6000
Dendritic Arbor by Field
**
Dorsal Contra
lpsi
C
Maximum Distance
6
1000
ce E
==t 0,
*E
E ‘5
Otl
z a
2
a0
G
.-
n
*
600
600
.- C
.g 400
200
0
Dorsal
Contra
lpsi
subjects t test, two-tailed; Fig. 6B). The dendritic extent of mo-
toneurons innervating hFl-treated muscles was 50% less in the
dorsal field compared to motoneurons innervating T-treated
muscles 0, < O.Ol), and 45% less in the contralateral field (p
< 0.05). Dendritic lengths within the ipsilateral field did not
differ significantly by treatment at the muscle (p > 0.05). A
modified Tukey post hoc test showed that the changes in the
dorsal and contralateral fields were significantly different from
the change in the ipsilateral field @ < 0.01 and p < 0.05, re-
spectively), and that the changes in the dorsal and contralateral
fields were not significantly different from one another. In other
words, there were nonuniform differences between fields, with
almost 50% less dendritic arbor length in the dorsal and contra-
lateral fields and no significant change in the ipsilateral field with
anti-androgen treatment at the muscle.
Maximum distance of labeled dendrites
To examine whether treatment at the muscle influenced CT-HRP
transport within dendrites, the maximum distance of labeled den-
drites from the cell body origin point was compared. A two-way
repeated-measures ANOVA of maximum distance per field by
treatment showed a significant main effect of treatment at the
muscle [F(l, 17) = 4.64; p < 0.051, a main effect of field [F(2,
17) = 18.99; p < O.OOOl], and no significant interaction of field
t
Legend
= TSide
El=
hFI Side
** = p<o.o1
* = p < 0.05
Figure
androgen stimulation to the BC/LA tar-
get muscles caused large changes in the
dendritic extent of SNB motoneurons.
A, Motoneurons innervating androgen-
deprived muscles had dendritic arbors
which were reduced by 44% overall
relative to those of motoneurons inner-
vating androgen-stimulated
Analysis of dendritic
within three regions of the transverse
spinal cord. Motoneurons
androgen-deprived
drite lengths which were reduced 50%
in the dorsal field, 45% in the contra-
lateral field, but did not differ signifi-
cantly in the ipsilateral field. C, The
maximum distance of CT-HRP-labelled
processes from the cell body origin
point differed only in the contralateral
field, and was 13% shorter for moto-
neurons innervating androgen-deprived
muscles relative to motoneurons inner-
vating androgen-treated muscles.
6. Thirty days of differential
muscles. B,
arbor lengths
innervating
muscles had den-
by treatment [F(2, 34) = 1.33; p > 0.051. Of the three fields
only the contralateral field showed a significant difference in the
maximum extent of dendrites from the origin point (Fig. 6C).
This distance was about 13% shorter in motoneurons innervating
hFl-treated muscles compared with motoneurons innervating
T-treated muscles. These results indicate that the differences in
dendritic lengths within each field were not due to altered ability
of CT-HRP to transport into dendrites with lateralized androgen
treatment at the muscle.
Discussion
Differential androgen treatment at the BC/LA muscle complex
caused substantial changes in the length of SNB motoneuron
dendrites approximately 8 cm away in the spinal cord. Several
different morphological measures were made in order to char-
acterize the nature of these changes, as well as to evaluate local
and systemic effects of the androgen capsule treatment.
Local and systemic effects of capsule treatment
As previously reported (Rand and Breedlove, 1992) the weights
of the lateral BCYLA muscle complexes within each animal dif-
fered significantly, and in the present study the hFl-treated mus-
cles weighed an average of 17% less than the T-treated muscles.
The difference in muscle weights was consistent and indicated
which innervated androgen-stimulated
estimated lengths were subtracted from those of the dendrites (see Materials and Methods). Scale bar, 400 pm.
muscles (B). Labeled axonal processes were identified by their characteristic “hook” shape (arrow), and their

Page 7

4414 Rand and Breedlove * Motoneuron Morphology and Androgen Stimulation of Target Muscles
the
stimulation
that this difference
actually has a slight anabolic
Breedlove, 1992). In the present study the induced
was often quite striking,
either side of the median
At the same time, the seminal
mental animals were not significantly
untreated castrated animals,
those of gonadally intact animals.
vesicle weights of the experimental
animals had very low levels of circulating
ated an experimental paradigm
androgen-deprived, and motoneurons
drogen-deprived BC/LA muscles while
er side innervated androgen-stimulated
gens entering circulation from the treatment
equal bilateral effects upon distant tissues such as the brain or
spinal cord. Because there are no known
nisms for retrograde androgen transport
differences in SNB morphology
gens acting at or near the BC/LA
successful
induction of a strong
muscles. We have previously
is not due to toxicity
effect upon the BC/LA
lateral gradient of androgen
to the BC/LA shown
by hFl, since hFI alone
(Rand and
asymmetry
in LA size on with
raphe (Fig. 2B).
large differences
vesicle weights
different
and were substantially
The greatly-reduced
group
of the experi-
from those of
smaller than
seminal
indicate that these
androgens.
both SNB nuclei were
on one side innervated
motoneurons
BC/LA muscles. Andro-
site would
This cre-
in which
an-
on the oth-
have had
or postulated
within
be mediated
muscle complex.
mecha-
cells, any lateral
by andro- would
CT-HRP
CT-HRP
reported
techniques
the important
by their projections.
androgens
onstrate large changes
ectomized
1986). However,
quent study from
changes in SNB dendritic
androgens might cause enhanced
within SNB motoneurons
aware of any direct evidence for this idea, and several lines of
evidence indicate that androgens
or transport: nevertheless, the present
minimize concerns that androgens
port of CT-HRP. By using local androgen
varied laterally all animals had nominally
of systemic androgens; the only difference
belled motoneurons innervated T-treated
However,
because
this issue is critically
tation of the present results, the rationale
given below.
Subsequent to the suggestion that androgens
HRP transport it has been shown that androgens
affect the rate of CT-HRP retrograde
site at the BC muscle to SNB somas. In the adult rat SNB system
CT-HRP first arrives at SNB cell bodies between
after injection, and all cell bodies are labeled
jection regardless of the presence or absence of systemic andro-
gens (Leslie et al., 1991). In the present
permitted to transport for 90-100
about X0 more hours for CT-HRP
drites after reaching the cell bodies. The capsules were removed
at the time of CT-HRP injection
during uptake of the retrograde
ruetrsurenlent
is a relatively
to be more sensitive than standard
for visualizing dendrites
advantage that it allows
The initial
on motoneuron
in dendritic
animals having 56% less arbor per cell (Kurz
using a different
the same laboratory
arbors, and speculated
of clerdritic
new retrograde
length
marker which
Golgi
has been
silver stain
and has
of neurons
(Wan et al., l982),
identification
report of the effect of systemic
morphology used CT-HRP
arbor lengths,
to den-
with gonad-
et al..
labeling technique
found
a subse-
modest only
that systemic
uptake or transport of CT-HRP
(Sasaki and Arnold, 1991). We are not
do not affect CT-HRP
study was designed
might affect retrograde
treatment
identical,
was whether
or hFI-treated
relevant to the interpre-
for using CT-HRP
uptake
to
trans-
only which
low levels
the la-
muscles.
is
might affect CT-
in fact do not
transport from the injection
X and IO hr
by 16 hr postin-
study CT-HRP was
hr after injection,
to distribute
allowing
the den- within
to curtail androgen
marker by the motoneuron
stimulation
ter-
minals and minimize
no difference
treatment
did not alter the number of motoneurons
or the number of motoneurons
tected.
Previous studies have shown
tween the presence of systemic androgens
HRP labeling in motoneuron
labeled motoneurons
gland of the Japanese quail were found to have equal dendritic
arbor lengths in both sexes, despite higher
in males (Seiwert and Adkins-Regan,
rat SNB system, studies using CT-HRP
dendrites extend maximally by the fourth postnatal
low, prepubertal androgen levels, and that when androgen
later rise SNB dendritic arbor lengths are reduced
the earlier time point (Sengelaub
et al., 1990). These data do not rule out an effect of androgens
on dendritic transport, but if such an effect does exist they argue
against its potential significance.
In the present study, analyses of the labeled
cated that androgens affected the length
dendrites rather than distribution
drites. The maximum distance of labeled dendrites from the cell
body origin point was not different
fields of the transverse spinal cord, and only 13% shorter in the
contralateral field of motoneurons
cles (Fig. 6C). In other words CT-HRP
distances into dendrites in two of three regions in the transverse
spinal cord, and the remaining
duced to explain the overall difference
Qualitatively, SNB motoneurons
had a lower density of dendritic
plane of the spinal cord, while
some of the dendrites were relatively
The small difference in the contralateral
was probably a modest change in the length of its most distal
dendrites.
Finally, the different proportional
dritic arbors within the three fields of the transverse plane sup-
port the idea that there were alterations
The ipsilateral dendritic field was not significantly
treatment at the muscles, whereas both the dorsal and contralat-
era1 fields showed large differences
respectively). It seems less likely that internal
be facilitated in two dendritic fields yet unaffected
field, and more likely that differential
length occurred as a result of androgen
This argument assumes that local, external
nificant influence during extension
bors which might account for the observed
arbors by field.
Taken together, previous reports and the present observations
show no effect of systemic androgens
no evidence for differential transport
neuron dendrites due to androgen
evidence instead indicates that SNB
change as a result of differential
muscles. The relative magnitude
study-nearly twice as much arbor per cell for SNB motoneu-
rons innervating androgen-treated
any possible local effects. There was also
of cells labelled
that local androgen
which took up CT-HRP
in which CT-HRP
in the number
at the muscle, showing
with CT-HRP by
treatment
could be de-
that there is no correlation
and the extent of CT-
dendrites. For example,
innervating the sexually
be-
CT-HRP-
dimorphic foam
levels of androgens
1990). In the developing
labeling report that SNB
week during
levels
compared
1986; Goldstein
to
and Arnold,
processes indi-
of SNB motoneuron
of CT-HRP within the den-
in the ipsilateral and dorsal
innervating hFl-treated mus-
equal was transported
region was not sufficiently
in dendritic
innervating hFl-treated
branches within
the overall lengths
unchanged
field indicates that there
re-
arbor lengths.
muscles
the transverse
of at least
(see Fig. 5).
changes in the labelled den-
in dendritic arbor length.
affected by
by treatment (50% and 45%.
transport would
in another
changes in dendritic
treatment
factors exert a sig-
or retraction of dendritic
differences
arbor
at the targets.
ar-
in the
on CT-HRP
of CT-HRP
treatment at distal targets. The
dendritic
androgen treatment at the target
of the effect in the present
transport and
within moto-
arbor lengths
muscles compared to moto-

Page 8

The Journal of Neuroscience, June 1995, 75(6) 4415
neurons
same as that reported
systemic androgen
fects for systemic and local androgen
the idea that the muscles are the primary
tion. It is also possible that these changes were facilitated
overall loss of systemic androgens;
at the level of the motoneurons
might have weakened synaptic
androgenic stimulation of the targets to more greatly
mechanisms of dendritic growth
each SNB motor nucleus overlap
innervating androgen-stimulated
able to maintain or compete for afferent innervation.
innervating androgen-depleted
in Kurz et al., 1986, a study which
manipulations. The observation
muscles-is about the during
the changes reported
androgen
gen may act primarily
arbors. These alterations
changes in the number of gap junctions
tons as well as reorganization
ioral consequences
therefore likely to be modulated
organization of SNB motoneurons
LA muscle size and contractile
The finding that steroid hormones
the morphology of innervating
unique. Despite wide-ranging
evoked by steroid treatment,
which possess steroid hormone
respond directly to hormones
the present study, central neurons can undergo
phology via steroids acting
influence” of steroid hormones
the distribution of steroid receptors suggests. This was always a
formal possibility, but our results emphasize
of steroid hormone action on the nervous system should care-
fully consider the potential mediating
ferent neurons.
copulation (Rose and Collins,
here compared
(Kurz et al., 1986) indicates
at the muscles to alter SNB
would be expected to cause extensive
between
of afferent synapses. The behav-
of androgen stimulation
by widespread
in addition
strength.
can act upon targets to alter
neurons
behavioral
the proportion
receptors and that can therefore
is small-less
198.5). The magnitude
to those seen with systemic
of
used
of similar ef-
supports
manipulations that andro-
dendritic administration
site of androgen’s ac-
by the
SNB motoneu-
that is, reduced
and their supraspinal
inputs and allowed
androgens
afferents
differential
influence
Dendrites
and motoneurons
of the muscles are
functional
to changes in BC/
re-
or retraction.
extensively,
muscles may have been better
from
is, to our knowledge,
changes
of central
that can be
neurons
Sninu .ri,-e
Previous
tion alters the size of SNB somas (Breedlove
Kurz et al., 1986). A recent report concluded
cles regulate the soma size of SNB motoneurons,
vation of SNB motoneurons
effects of systemic androgens
study did not evaluate whether
recovery as a result of the androgen
interpret because androgens
(Yu and Srinivasan, I98 1).
In the present study systemic androgen
low and soma size did not differ by treatment
using either a within-subjects
a between-subjects analysis of CT-HRP-filled
sults indicate that androgens
toneurons to affect soma size. In contrast,
in dendritic arbor lengths with androgen
implies that SNB dendritic
produced by the muscles in response to androgen
this is the case, the muscle-produced
act independently of androgen’s
studies have found that systemic androgen manipula-
than 1%. If. as in
changes in mor-
and Arnold,
that BC/LA
using reinner-
19X1;
mus- upon targets, then the “sphere
is potentially
of
much larger than
to different target muscles to assess
(Araki et al., I99 I ). However, this
there was differential
treatment
are known to affect reinnervation
that future studies
functional
and is difficult to
effects of targets on af-
levels were uniformly
at the muscles,
analysis of thionin-stained cells or
somas. These re-
upon SNB mo-
the large difference
treatment at the muscles
might act directly
arbor size is mediated by some factor
stimulation.
factor might
If
neurotrophic
effects on SNB soma size.
Using axotomy and reinnervation
shown an association
size of their innervating
Yawo, 19X7), and in one developmental
were reduced or enlarged
nipulation of the target tissue size (Voyvodic,
servations indicate that targets normally
some type of trophic
adult SNB system this trophic
drogens. Assuming this is true, it is not clear
strength of the neurotrophic
changes in the muscles’ size. Data from gonadally
which were implanted with dexamethasone
BC/LA muscles indicate
lengths, despite an I I% reduction
sone-treated BC/LAs (Rand and Breedlove,
vations). It is therefore possible that SNB dendritic
might be mediated by androgens
SNB dendrites make gap junctions
somas and dendrites, and the number of these gap junctions
be regulated by systemic androgens
Gap junctions serve to electrotonically
be important for bilateral coordination
of targets, several reports have
between target cells and the dendritic
neurons (Sumner
study dendritic
commensurate with experimental
arbor
1971;
arbors
ma-
and Watson,
1989). These ob-
supply neurons with
support. Our results suggest that in the
support can be modulated by an-
whether
associated with
intact animals
capsules next to the
in SNB dendritic
in weight of the dexametha-
unpublished
the
signal is exclusively
no changes arbor
obser-
arbor lengths
of target size.
SNB cell
independently
with contralateral
can
(Matsumoto
couple cells, which
of the BC/LA
et al., 1988).
may
muscles
References
Araki I. Harada Y, Kuno M ( 199 I ) Target-dependent hormonal control
of neuron size in the rat spinal nucleus of the hulbocavernosus. J
Neurosci 11:3025-3033.
Breedlove SM, Arnold AP (1980) Hormone accumulation in a sexually
dimorphic motor nucleus in the rat spinal cord. Science 2105644
566.
Breedlove SM. Arnold AP (198 I ) Sexually dimorphic motor nucleus
in the rat lumbar spinal cord: response to adult hormone manipula-
tion. absence in androgen-insensitive
Breedlove SM, Arnold AP (19X3) Hormonal control of a developing
neuromuscular system. 11. Sensitive periods for the androgen induced
masculinization of the rat spinal nucleus of the bulbocnvernosus. J
Neurosci 3:424432.
Cihak R, Gutmann E, Hanzlikova V (1970) Involution and liormone-
induced persistence of the muscle sphincter (levator) ani in female
rats. J Anat 106:93-l IO.
Forger NG, Breedlove SM (1987) Seasonal variation in mammalian
striated muscle mass and motoneuron morphology. J Neurobiol
155-165.
Goldstein LA, Kurz EM, Sengelaub DR (1990) Androgen regulation
of dendritic growth and retraction in the development of a sexually
dimorphic spinal nucleus. J Neurosci 10:935-946.
Hayes KJ (1965) The so-called “levator ani” of the rat. Acta Endo-
crinol 48:3377347.
Joubert Y. Tobin C (1989) Satellite cell proliferation
the number of myonuclei induced by testosterone in the levator ani
muscle of the adult female rat. Dev Biol I3 1:550-557.
Jung I, Baulieu E-E (I 972) Testosterone cytosol “receptor”
levator ani muscle. Nature New Biol 237:24-26.
Konigsmark BW (1970) Methods for the counting of neurons. In: Con-
temporary research methods in neuroanatomy (Nauta WJH, Ebbeson
SOE, eds), pp 315-340. New York: Springer.
Kurz EM, Sengelaub DR, Arnold AP (1986) Androgens regulate the
dendritic length of mammalian motoneurons in adulthood. Science
232:X-398.
Leedy MC, Beattie MS, Bresnahan JC (1987) Testosterone-induced
plasticity of synaptic inputs to adult mammalian motoneurons. Brain
Res 424:386-390.
Leslie ML, Forger NG, Breedlove SM (1991) Does androgen affect
axonal transport of cholera toxin HRP in spinal motoneurons’? Neu-
rosci Lett 126: 199-202.
Matsumoto A, Arnold AP Zampighi GA, Micevych PE ( 1988) Andro-
rats. Brain Res 225:207-305.
18:
and increase in
in the rat

Page 9

4416 Rand and Breedlove * Motoneuron Morphology and Androgen Stimulation of Target Muscles
genie
cord.
regulation
J Neurosci
M
of gap junctions
8:4177-4183.
Tetramethyl
a non-carcinogenic
for visualizing
Cytochem 26: 106-I
EP Breedlove
of the bulbocavernosus.
EJ, Nordeen KW,
prevent normally
nucleus. Science
MN, Breedlove
bocavernosus and levator
Rose RD, Collins WF
for synchronized
373-377.
Sasaki M, Arnold AP
of motoneurons in the spinal
struction after intracellular
Comp Neurol 308: 1 l-27.
Schroder HD (1980)
pelvic muscles of the male
Seiwert CM, Adkins-Regan
between motoneurons in rat spinal
Mesulam
neurohistochemistry:
superior
Histochem
Monaghan
nucleus
Nordeen
gens
spinal
Rand
(1978) benzidine for
blue
afferents
horseradish
reaction-product
peroxidase
with
sensitivity neural and efferents. J
17.
SM (1991) Brain sites projecting
Neurol
Arnold
in a sexually
to the spinal
J Comp 307:370-374.
AP (1985) Sengelaub
occurring
229:671-673.
(1992)
ani size. J Neurobiol
(1985) Crossing
activation of penile
DR, Andro-
cell death dimorphic
SM Androgen locally regulates rat bul-
23: 17-30.
may dendrites
motoneurons.
be a substrate
Brain Res 337:
(1991) Androgenic
nucleus
iontophoresis
regulation
of the bulbocavernosus:
of horseradish
of dendritic trees
recon-
peroxidase. J
Organization of the motoneurons
rat. J Comp
E (1990) Peripheral
innervating the
Neurol 192:567-587.
sexual dimorphism
without
romuscular
Sengelaub
jections
6:1613-1620.
Sumner
dendritic
2231273-275.
Venable
deprived,
119:271-302.
Voyvodic
in the rat superior
Wan X-CS,
tecture
nucleus
with cholera
Yawo H (1987)
cervical
7:3703-371
Yu WA, Srinivasan
drotestosterone
Neurol
corresponding
system.
DR. Arnold
in a sexually
CNS
Neurosci
AP (1986)
dimorphic
dimorphism in a hormone-sensitive
16:743.
Development
rat spinal
neu-
Sot Abstr
and loss of early
nucleus.
pro-
J Neurosci
BEH, Watson WE (1971) Retraction
of adult
and expansion
in rsi\lo. Nature
of the
tree of motor neurones rats induced
JH (I 966)
and testosterone-stimulated
Morphology of the cells of normal,
ani muscles.
testosterone-
Am J Anat levator
JT (1989) Peripheral
cervical
target regulation
J Neurosci
JO, Liu
of dendritic
9: 1997-2010.
CN (1982)
dendrites
of horseradish
geometry
ganglion.
Gonatas
and commissural
by conjugates
Neurol 78:1677175.
in the dendritic
cells following
Trojanowski
of the extranuclear
neurons
toxin.
Changes
ganglion
I.
JQ, Cytoarchi-
of hypoglossal
peroxidase as revealed
Exp
geometry of mouse
axotomy.
superior
J Neurosci postganglionic
R (1981)
on regeneration
Effect of testosterone
of the hypoglossal
and 5 alpha
nerve
dihy-
in rats. Exp
71:431435.