Development of Electrophysiological and Morphological Diversity
in Autonomic Neurons
REBECCA L. ANDERSON, PHILLIP JOBLING, AND IAN L. GIBBINS
Centre for Neuroscience and Department of Anatomy and Histology, Flinders University, Adelaide, SA 5001, Australia
Received 8 March 2001; accepted in final form 1 June 2001
Neurons vary widely in their expression of ion channels, especially voltage-dependent K⫹ channels and Ca2⫹-dependent K⫹
channels. These channels are a fundamental determinant of the
firing properties of neurons and their responses to synaptic inputs
(Hille 1992). Integration of convergent synaptic inputs also is
dependent on interactions between the distribution of ion channels
and the dendritic morphology of the neurons. The generation of
neuronal diversity clearly requires the coordinated development of
these features (Dryer 1994, 1998; Ribera and Spitzer 1992).
However, the relationship between differential ion channel expression and dendritic morphology during the development of
mature neuronal phenotypes is not well known.
Autonomic pathways comprise one of the primary motor
outputs of the nervous system and contain more final motor
neurons than any other pathway. In humans, there are more
than 10 million final motor neurons in sympathetic pathways
alone (Gibbins 1990). Compared with somatic final motor
neurons, autonomic neurons show a great diversity of phenotypic characteristics, such as their neuropeptide content, electrical properties, morphology, and synaptic connectivity. In
addition, autonomic neurons found in specific functional pathways often express precise combinations of these phenotypic
characteristics (Adams and Harper 1995; Andrews et al. 1996;
Chiba and Tanaka 1998; Dryer 1994; Gibbins 1995; Jobling
and Gibbins 1999; Morris et al. 1997–1999; Smith 1994). The
celiac ganglion of guinea pigs provides a striking example of
this phenomenon. Here, vasomotor neurons can be distinguished from neurons projecting to the enteric plexuses by
their location, the size of their dendritic fields, their neuropeptide content, the potassium channels they express, and the
origins and number of their synaptic inputs (Boyd et al. 1996;
Cassell and McLachlan 1987; Cassell et al. 1986; Costa and
Furness 1984; Davies et al. 1999; Gibbins et al. 1999; Keast et
al. 1993; Lindh et al. 1986; Macrae et al. 1986; McLachlan and
Meckler 1989; Meckler and McLachlan 1988) (Table 1).
Although autonomic neurons have been used to study many
different aspects of neuronal differentiation (Dryer 1994, 1998;
Dryer and Chiappinelli 1985; Hirst and McLachlan 1984;
McFarlane and Cooper 1992, 1993; Nerbonne and Gurney
1989; Phelan et al. 1997; Rubin 1985a– c), few studies have
examined the development of phenotypic diversity within
functionally identified pathways (Cameron and Dryer 2000;
Stofer and Horn 1990, 1993). Indeed, most studies of neuronal
development have investigated only a single class of neurons.
In principle, the generation of different neuronal phenotypes
from a common precursor pool could occur by the sequential
Address for reprint requests: R. L. Anderson, Centre for Neuroscience and
Dept. of Anatomy and Histology, Flinders University, GPO Box 2100, Adelaide, SA 5001, Australia (E-mail: [email protected]).
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked ‘‘advertisement’’
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
INTRODUCTION
www.jn.org
0022-3077/01 $5.00 Copyright © 2001 The American Physiological Society
1237
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
Anderson, Rebecca L., Phillip Jobling, and Ian L. Gibbins. Development of electrophysiological and morphological diversity in
autonomic neurons. J Neurophysiol 86: 1237–1251, 2001. The generation of neuronal diversity requires the coordinated development of
differential patterns of ion channel expression along with characteristic
differences in dendritic geometry, but the relations between these phenotypic features are not well known. We have used a combination of
intracellular recordings, morphological analysis of dye-filled neurons,
and stereological analysis of immunohistochemically labeled sections to
investigate the development of characteristic electrical and morphological
properties of functionally distinct populations of sympathetic neurons that
project from the celiac ganglion to the splanchnic vasculature or the
gastrointestinal tract of guinea pigs. At early fetal stages, neurons were
significantly more depolarized at rest compared with neurons at later
stages, and they generally fired only a single action potential. By mid fetal
stages, rapidly and slowly adapting neurons could be distinguished with
a topographic distribution matching that found in adult ganglia. Most
rapidly adapting neurons (phasic neurons) at this age had a long afterhyperpolarization (LAH) characteristic of mature vasomotor neurons and
were preferentially located in the lateral poles of the ganglion, where
most neurons contained neuropeptide Y. Most early and mid fetal neurons showed a weak M current, which was later expressed only by
rapidly-adapting and LAH neurons. Two different A currents were
present in a subset of early fetal neurons and may indicate neurons
destined to develop a slowly adapting phenotype (tonic neurons). The
size of neuronal cell bodies increased at a similar rate throughout development regardless of their electrical or neurochemical phenotype or their
topographical location. In contrast, the rate of dendritic growth of neurons
in medial regions of the ganglion was significantly higher than that of
neurons in lateral regions. The apparent cell capacitance was highly
correlated with the surface area of the soma but not the dendritic tree of
the developing neurons. These results demonstrate that the well-defined
functional populations of neurons in the celiac ganglion develop their
characteristic electrophysiological and morphological properties during
early fetal stages of development. This is after the neuronal populations
can be recognized by their neurochemical and topographical characteristics but long before the neurons have finished growing. Our data
provide strong circumstantial evidence that the development of the full
phenotype of different functional classes of autonomic final motor neurons is a multi-step process likely to involve a regulated sequence of
trophic interactions.
1238
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
1. Phenotypic characteristics of major classes
of functionally identified neurons in mature guinea pig
coeliac ganglion
TABLE
Main target
Location in ganglion
Neuropeptide content
Dendritic field size
IsAHP
IM
IA
Slow IA
Firing properties
Peripheral synaptic
inputs
LAH Neurons
Tonic Neurons
Blood vessels
Lateral
NPY
Small
Present
Present
Small, inactive at rest
Absent
Phasic (rapidly-adapting)
Enteric plexuses
Medial
Som or none known
Large
Absent
Usually absent
Large, active at rest
Present in 40%
Tonic (slowly-adapting)
Absent
Present
Electrophysiology
Celiac ganglia, their nerve trunks, and surrounding tissues (aorta, celiac artery and adrenal glands) were removed and placed into a HEPES-buffered balanced salt solution
containing (in mM) 146 NaCl, 4.7 KCl, 0.6 MgSO4, 1.6 NaHCO3,
0.13 NaH2PO4, 2.5 CaCl2, 7.8 glucose, and 20 HEPES, buffered to pH
7.3 and gassed with 100% O2. Ganglia were pinned to the base of a
recording chamber (Medical System, Greenvale, NY) lined with silicone elastomer (Sylgard, Dow Corning, Midland, MI). During electrophysiological recordings, ganglia were maintained at 35°C and
superfused with HEPES balanced salt solution at 2.5 ml/min.
At early fetal stages (F30 –F35), poorly developed connective tissue
did not allow the celiac ganglion to be pinned tightly in the recording
chamber. Instead the ganglion was stabilized by leaving it attached to
the abdominal aorta, which was slit longitudinally along its dorsal
surface and the reflected corners pinned down. In addition, early fetal
neurons were small with little cytoplasm (cross-sectional area of soma
⬃150 ␮m2) (Anderson et al. 2001). Both of these factors affected the
length of time for which impalements could be held as has been
reported in other preparations of developing sympathetic ganglia
(Dryer and Chiappinelli 1985; Hirst and McLachlan 1984). Preparations from the earliest fetal stages examined (F30 –F32) were only
viable for ⬃2 h at 35°C, after which time the ganglion and connective
tissue began to deteriorate. At later stages (F38⫹), preparations were
viable for ⱕ8 h at 35°C (longest time attempted) and impalements
were routinely held for ⬎20 min.
TISSUE PREPARATION.
acquisition of characteristics from a basal embryonic phenotype, such that one mature phenotype was derived from another. Alternatively, each mature phenotype could develop
directly from a specific subset of precursors. The resolution of
this question has been hampered by the difficulty of identifying
different pools of precursor neurons prior to their differentiation. In this study, we have tackled this question by taking
advantage of the unique organization of the guinea pig celiac
ganglion, which allows us to follow the development of phenotypically diverse populations of neurons that innervate distinct target tissues. We have investigated the differentiation of
two major phenotypic characteristics, the differential expression of ion channels, and dendritic morphology in neurons
whose functional pools can be recognized from an early developmental stage simply on the basis of their location. To do
this, we used intracellular electrophysiological recording techniques, combined with dye-filling, multiple-labeling immunohistochemistry, and confocal microscopy. We have found that
many of the electrophysiological characteristics of the main
functional classes of neurons develop directly from undifferentiated precursors and can be distinguished from each other
long before the neurons finish growing. This suggests that the
differentiation of these two phenotypic characteristics is likely
to be independently regulated.
METHODS
Pregnant guinea pigs, fetuses, neonates (P1–P13) and nonpregnant
adults (⬎240 g; Cavia porcellus, Hartley/IMVS strain) were given a
lethal dose of sodium pentobarbitone (Nembutal, Bomac Laboratories, Asquith, Australia; 200 mg/kg ip). Nonpregnant adult guinea pigs
used in the stereological analysis of neuropil (see following text) were
killed by stunning and exsanguination. Late-stage fetuses also were
exsanguinated after removal from their extra-embryonic membranes.
Fetuses were then weighed and placed in a balanced salt solution (see
following text). All procedures were approved by the Animal Welfare
Committee of the Flinders University of South Australia.
Guinea pigs have a relatively long and variable gestation period of
between 55 and 75 days (Matsumoto et al. 1993; Weir 1974). Embryogenesis occurs during the first 30 days of gestation, while the
remainder of the gestational period involves fetal growth (Scott 1937).
Since guinea pigs undergo postpartum estrus within hours of giving
birth (Stockard and Papanicolou 1917), the day of birth of the previous litter is also day 0 of the following litter. Fetuses were obtained
from pregnant guinea pigs during three arbitrary stages of develop-
INTRACELLULAR RECORDINGS OF SYMPATHETIC NEURONS. Neurons were impaled using high-resistance glass microelectrodes (80 –200
M⍀) pulled on a Flaming-Brown puller (Sutter Instrument, Novarto, CA)
and filled with 0.5 M KCl. Electrical properties were determined using
bridge mode, discontinuous current clamp (DCC), or single electrode
voltage clamp (SEVC) using either an Axoprobe-1A or an Axoclamp-2B
amplifier (Axon Instruments, Union City, CA). Voltage or current
records were digitized at 1–5 kHz using Spike2 (version 3.01) and Signal
software (version 1.72; Cambridge Electronic Design, Cambridge, UK)
on a PC running Windows NT, or Chart/Scope software (version 3.5,
MacLab, ADI Instruments, Castle Hill, NSW, Australia) on a Power
Macintosh computer (Apple Computers, Cupertino, CA). During DCC
and SEVC, the headstage was continuously monitored and the cycling
frequency adjusted to minimize the effects of electrode capacitance. The
cycling frequency was 1.0 –2.0 kHz for DCC and 2.0 –3.5 kHz for SEVC.
Digitized data were analyzed using Igor Pro (version 3.14, WaveMetrics,
Lake Oswego, OR).
Resting membrane potential (RMP) was determined by measuring
the difference between the potentials immediately before and after
withdrawal of the microelectrode from the cell. Measurements of
input resistance (Rin) and the major input time constant were made by
injecting small hyperpolarizing current pulses (0.01– 0.1 nA, 200 –250
ms duration) through the recording electrode. Averages of 20 current
steps were routinely used. The time constant of the cell (␶, ms) was
determined by fitting a single exponential to the onset of the voltage
response to the current pulse between 20 and 80% of its final amplitude. Capacitance was derived from the time constant divided by the
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
See text for references. LAH, long after hyperpolarization; NPY, neuropeptide Y; Som, somatostatin.
J Neurophysiol • VOL
ment as previously described (Anderson et al. 2001): early fetal
[F30 –F35; weight range 1–7 g, mean 3.4 ⫾ 0.3 (SE) g, n ⫽ 30], mid
fetal (F36 –F45; weight range 10.2– 48.9 g, mean 20.7 ⫾ 1.9 g, n ⫽
21), and late fetal (F46⫹; weight range 37–79 g, mean 50.4 ⫾ 7.5 g,
n ⫽ 5). The early fetal stage of development in guinea pigs is
approximately equivalent to the first postnatal week in rats and mice
(Butler and Juurlink 1987). The weight of neonatal guinea pigs
(P0 –P13) used in this study ranged from 93 to 191 g (mean 126.0 ⫾
8.4 g, n ⫽ 13), while nonpregnant adults ranged from 240 to 329 g
(mean 286.0 ⫾ 18.2, n ⫽ 5). Where possible, our data were analyzed
using the log of the weight since there was an exponential increase in
weight with increasing age.
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
Neuronal morphology
RELATIVE AREA OF NEUROPIL IN TOPOGRAPHICALLY DISTINCT
REGIONS OF THE CELIAC GANGLION. Celiac ganglia were removed
from embryos (Carnegie stages 20 –23), fetuses, neonates, and adults,
fixed in Zamboni’s fixative (0.2% picric acid and 2% formaldehyde in
0.1 M phosphate buffer, pH 7.0), and processed for multiple-labeling
immunohistochemistry (Anderson et al. 2001). Ganglia were dehydrated in ethanol (EtOH), cleared in DMSO, washed in 100% EtOH,
and vacuum infiltrated at 46°C for ⱖ30 min in polyethylene glycol
(PEG; MW 1000), before being embedded in PEG (MW 1450) in
small cryomolds. Sections, 10- to 20-␮m thick, were cut on a standard
rotary microtome and placed into phosphate-buffered saline (PBS).
Excess PBS was removed and the sections placed in 10% normal
donkey serum (NDS) for 30 min. Sections were incubated in 10%
NDS and primary antisera at room temperature for 48 –72 h. Labeling
for neuropeptide Y (NPY) was used to identify vasomotor neurons,
and labeling for somatostatin (Som) was used to identify neurons
projecting to the enteric plexuses; labeling for tyrosine hydroxylase
(TH) was used as an internal labeling control because nearly all celiac
ganglion neurons contain TH regardless of their peptide content
(Anderson et al. 2001; Costa and Furness 1984). Primary antibodies
used were: sheep anti-NPY (Oliver/Blessing E2210/2; 1:1000) or
rabbit anti-NPY (Incstar, Stillwater, MN, No. 550212; 1:1200), monoclonal mouse anti-Som (MRC Regulatory Peptide Group, Vancouver,
Canada; code Soma S8; 1:1200) or rabbit anti-Som (Incstar; 1:100),
and in some cases mouse anti-TH (Incstar, No. 105440, 1:1200) or
rabbit anti-TH (Dr. J. Thibault, AS2–512, 1:2000). After washing in
PBS, secondary antibodies were applied for ⱖ2 h. Species-specific
secondary antibodies (IgG) were raised in donkeys and conjugated
with dicholortriazinyl amino fluorescein (DTAF), fluorescein isothiocyanate (FITC) or the indocarbocyanin dyes Cy3 or Cy5. All secondary antibodies were obtained from Jackson ImmunoResearch Laboratories, West Grove, PA. After further washing, sections were
mounted on glass slides in carbonate-buffered glycerol (pH 8.6), and
the coverslips were sealed using clear nail polish.
J Neurophysiol • VOL
Sections were examined using conventional wide-field fluorescence or
confocal microscopy. For conventional microscopy, images were collected using an Olympus AX70 microscope (Olympus, Tokyo, Japan)
fitted with a Hamamatsu Orca cooled CCD camera (Hamamatsu Photonics, KK, Japan) and connected to a PowerMac G3 (Apple Computers)
running IPLab Spectrum (version 3.2, Scanalytics, Fairfax, VA). Confocal microscopy was done using a BioRad MRC-1024 scanning laser
confocal microscope (BioRad, Hemel Hempstead, UK) with a krypton/
argon laser source fitted to an Olympus AX70 epifluorescence microscope and running under LaserSharp software (version 3.2, BioRad).
A stereological point-counting method (Howard and Reed 1998)
was used to quantify changes in the proportional area of ganglion
occupied by neuropil in lateral and medial regions of the developing
celiac ganglion. Digital images of sections labeled for immunoreactivity to NPY, Som, and, in some cases, TH, were overlaid with a
25–35 point grid using Adobe Photoshop software (version 5.1,
Adobe Systems, Mountain View, CA). Points intersecting with neurons, nonneuronal tissue, and neuropil were scored. Up to four samples from medial or lateral locations were averaged in each animal
(see also Anderson et al. 2001).
NEUROBIOTIN-FILLED NEURONS. During some intracellular impalements, Neurobiotin (0.5% wt/vol in 0.5 M KCl; Vector, Burlingame,
CA) was included in the electrode filling solution so that neurons
could be visualized after the completion of electrophysiological experiments. The location of neurons in the bilobed celiac ganglion was
recorded as either in the medial two-thirds or in the lateral third of a
lobe. At the completion of the electrophysiological recordings, ganglia were fixed in Zamboni’s fixative for 24 –72 h and processed as
whole mounts for multi-labeling immunohistochemistry as previously
described (Gibbins et al. 1999; Jobling and Gibbins 1999). Briefly,
picric acid was removed by washing in 80% EtOH before the tissue
was further dehydrated in 100% EtOH and permeabilized in DMSO
for 1–3 h. Tissue was then rehydrated through 80 and 50% EtOH
before being washed in PBS (pH 7.0). Primary antisera for NPY and
Som (as in the preceding text at twice the concentrations used for
sections) were then applied for 48 –72 h. After extensive washing in
PBS, whole mounts were incubated in secondary antibodies overnight. Species-specific secondary antibodies (IgG), raised in donkeys
and conjugated with DTAF, FITC, or Cy3 (see preceding text) were
used to detect immunoreactivity to NPY and Som. Streptavidin conjugated to Cy5 (Jackson Immunoresearch Laboratories) was used to
detect Neurobiotin-filled cells. After 2– 4 h washing in PBS, ganglia
were mounted on glass slides in carbonate-buffered glycerol (pH 8.6).
Scanning laser confocal microscopy was used to analyze dye-filled
neurons. Neurons first were assessed for their immunoreactivity to NPY
or Som. To ensure antibodies had penetrated sufficiently through the
whole ganglion, dye-filled neurons were only scored as lacking neuropeptides if at least some adjacent neurons showed positive immunoreactivity. Dye-filled neurons were only included in the morphological
analysis if individual dendrites were easily distinguishable, if they were
not obscured by adjacent dye-filled neurons, and if the axon could be
identified and followed to the edge of the ganglion. A low-magnification
confocal through-focus series, which included all dendrites, was taken of
each filled neuron with optical sections separated by 0.5–2.0 ␮m. A
confocal through-focus series also was taken of the neuronal cell body at
higher magnification using low gain settings to confirm that only one
neuron had been filled and to provide a more precise measure of the
cross-sectional and surface area of the soma. A single two-dimensional
(2D) maximum-intensity projection image was generated from each
confocal series using either Lasersharp or National Institutes of Health
Image (NIH) software (version 1.61, Bethesda, MD). Measurements of
cross-sectional area of the neuronal cell body (␮m2), number of primary
dendrites (ⱖ1 cell body diameter in length), and the total dendritic length
(␮m) were taken. Confocal through-focus series also were reconstructed
on a PC using VoxBlast software for Windows (version 3.0, VayTek,
Iowa City, IA). Threshold was optimized either for dendrites (low-
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
Rin. Steady-state current-voltage (I-V) curves were generated by measuring the average response during the last 10 –30 ms of a 200- to
250-ms current step. Neurons were classified as phasic or tonic on the
basis of their firing properties in response to 200- to 250-ms depolarizing current steps (see Cassell et al. 1986; Keast et al. 1993). Phasic
neurons were further discriminated as long afterhyperpolarizing
(LAH) or not on the basis of the presence or absence of a prolonged
after hyperpolarization (AHP; Cassell and McLachlan 1987). For
measurement of action potential (AP) amplitudes, recordings were
made in bridge mode. Single APs were elicited by brief (10 –20 ms)
injections of depolarizing current from rest. The maximum amplitude
(mV) and half-width (ms) of the AP were measured together with the
maximum amplitude (mV) and duration of the afterhyperpolarization
(ms; measured from the point at which the cell passed RMP during
repolarization after the action potential until it returned to RMP).
For measurements of IAHP and IsAHP in SEVC, brief (10 –20 ms)
suprathreshold depolarizing voltage steps resulted in a single “action
current” that corresponds to an unclamped action potential (Cassell et al.
1986; Jobling et al. 1993; Wang and McKinnon 1995). Peak amplitudes
of IAHP and IsAHP were measured 10 ms after the action current. The time
constant of the IAHP was determined by fitting a single exponential
between 80 and 20% of the curve (Cassell and McLachlan 1987).
DRUG APPLICATIONS. The action potentials of developing neurons
either are initially dependent on Ca2⫹ before becoming Na⫹ dependent
or they are Na⫹ dependent from the onset of excitability (Spitzer 1991).
The ionic dependence of early fetal neurons was examined using tetrodotoxin (TTX, 1 ␮M; Alamone Labs, Jerusalem, Israel) to block Na⫹dependent channels and Cd2⫹ (300 ␮M; ICN Biomedicals, Costa Mesa,
CA) to block Ca2⫹-dependent channels (Adams and Harper 1995; Davies
et al. 1999). Solutions were changed by switching the perfusion line.
1239
1240
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
magnification series) or for the cell soma (high-magnification series), the
surface area calculated and images of three-dimensional reconstructions
rendered. The brightness and contrast of images was adjusted using
Adobe Photoshop software.
TESTS OF MORPHOLOGICAL MEASUREMENTS AFTER DIFFERENT
FIXATION AND MOUNTING TECHNIQUES. To test for any morpho-
RESULTS
Development of electrical properties of celiac ganglion
neurons
PASSIVE MEMBRANE PROPERTIES. Intracellular recordings were
made from 177 neurons in 63 preparations of celiac ganglia
from fetal, neonatal, and adult guinea pigs. The RMP of celiac
FIG. 1. Relationships between increasing body weight of the animal and passive membrane properties. Weight (in g) is shown on a
log scale. Approximate stages of development are indicated (F, fetal; Neo, neonate). The number of neurons (n), R2 value, and significance
of the regression (*P ⬍ 0.05) are indicated. Regression lines are shown with 95% confidence limits. A: the resting membrane potential
(RMP, mV) of neurons became significantly more hyperpolarized during development (R2 ⫽ 0.42, F(1,127) ⫽ 93.3, P ⬍ 0.0001). B: there
was no significant change in input resistance (Rin, M⍀) during development (R2 ⫽ 0.01, F(1,162) ⫽ 1.5, P ⫽ 0.2). C: the major cell input
time constant (␶, ms) increased significantly during development (R2 ⫽ 0.12, F(1,116) ⫽ 15.2, P ⬍ 0.0001). D: the apparent cell
capacitance (derived from measured Rin and ␶) significantly increased during development (R2 ⫽ 0.16, F(1,116) ⫽ 22.7, P ⬍ 0.0001).
J Neurophysiol • VOL
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
logical changes that may accompany tissue processing of dye filled
neurons, a conjugate of Dextran (10,000 MW), tetramethylrhodamine
and biotin (“Mini-Ruby”; 20 ␮l in 2% 0.5 M KCl; Molecular Probes,
Eugene, OR) was used to fill neurons in celiac ganglia from two adult
guinea pigs. The ganglia were then mounted on glass slides in the
same solution used during the experiments, and the coverslips were
held in place using nail polish. A confocal through-focus series of
each dye-filled neuron was captured before ganglia were removed
from the slides and fixed in Zamboni’s fixative. Following fixation,
ganglia were processed as for other experiments before they were
re-mounted in PBS. Then a second confocal through-focus series was
taken of each dye-filled neuron. Finally, ganglia were re-mounted in
buffered glycerol and left overnight before a third confocal throughfocus series of each dye-filled neuron was captured. A single 2D
maximum-intensity projection image was generated from each confocal series, and NIH image software was used to measure the
cross-sectional cell body area and total dendritic length. Three neurons sufficiently well filled for morphological analysis were followed
throughout all the steps. None of these neurons underwent any significant shrinkage or any other morphological deformations with the
fixation, processing, and mounting techniques used here.
STATISTICAL ANALYSIS. Development changes were analyzed with
least-squares linear regression, with log-transformed weight used as a
measure of developmental age. Means were compared with t-tests or
multivariate ANOVA, while medians of strongly skewed data were
compared with Mann-Whitney U tests. Frequency data were analyzed
using ␹2 tests. All analyses were done with SPSS for Windows (version
9, SPSS, Chicago, IL). Data are presented as untransformed means ⫾ SE,
with n values referring to the number of neurons unless otherwise stated.
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
all, the mean input resistance was 119.0 ⫾ 6.0 M⍀ (n ⫽ 169).
The major time constant (␶) increased significantly during
development, from ⬃7 ms at early fetal stages to ⱖ11 ms at
later stages (Fig. 1C). As a consequence, the apparent cell
capacitance also showed a significant increase with age from
⬃60 pF at early fetal stages to ⱖ100 pF at subsequent stages
(Fig. 1D).
ACTION POTENTIAL CHARACTERISTICS AND FIRING PROPERTIES.
The peak amplitude of the AP, the potential at which this peak
was reached, and the AP half-width were measured in neurons
that generated an AP in response to a 10- to 20-ms depolarizing
step. Although the peak amplitude of the AP increased significantly during development (Fig. 2A), the potential at which
this was reached (between 10 and 30 mV) did not change
significantly (R2 ⫽ 0.03, F(1,61) ⫽ 1.61, P ⫽ 0.2). Thus the
increase in peak AP amplitude is likely to reflect the fact that
FIG. 2. Relationships between increasing body weight of the animal and characteristics of the action potentials (APs) and
afterhyperpolarizations (AHP). Weight (in g) is shown on a log scale. Approximate stages of development are indicated (F, fetal;
Neo, neonate). The number of neurons (n), R2 value, and significance of the regression (*P ⬍ 0.05) are indicated. Regression lines
are shown with 95% confidence limits. A: the peak amplitude of the AP (mV) increased significantly during development (R2 ⫽
0.46, F(1,86) ⫽ 72.2, P ⬍ 0.001). B: there was a significant decrease in the duration of the AP (AP half-width, ms) during
development (R2 ⫽ 0.14, F(1,38) ⫽ 5.9, P ⫽ 0.02). C: during development, the amplitude of the AHP increased significantly (R2 ⫽
0.42, F(1,74) ⫽ 52.6, P ⬍ 0.0001). When analyzed separately, both tonic (R2 ⫽ 0.20, F(1,23) ⫽ 5.8, P ⫽ 0.025) and long
afterhyperpolarizing (LAH) neurons (R2 ⫽ 0.58, F(1,31) ⫽ 42.1, P ⬍ 0.0001) showed significant increases in AHP amplitude during
development. The regression line is shown for the total neuronal population. D: the duration of the AHP (ms) significantly increased
during development in tonic neurons (R2 ⫽ 0.22, F(1,22) ⫽ 6.3, P ⫽ 0.02), but significantly decreased in LAH neurons (R2 ⫽ 0.23,
F(1,28) ⫽ 8.6, P ⫽ 0.007).
J Neurophysiol • VOL
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
ganglion neurons became significantly more negative during
development (Fig. 1A). At early fetal stages (F30 –F35), the
majority of neurons had RMPs around ⫺35 mV while those
from neonates had RMPs around ⫺55 mV. In previous studies
of developing sympathetic ganglia using intracellular recording
techniques, neurons with RMPs outside published ranges for
mature sympathetic neurons were not considered for analysis
(e.g., Dryer and Chiappinelli 1985; Hirst and McLachlan
1984). However, in this study, half of the early fetal neurons
with RMPs around ⫺35 mV had input resistances ⱖ100 M⍀
(see following text; Fig. 1, A and B), suggesting that they were
unlikely to have been damaged significantly during intracellular impalements. Therefore we have included immature neurons with RMPs less negative than ⫺55 mV in the analyses
that follow.
In contrast to RMP, the input resistance (Rin) of neurons did
not change significantly during development (Fig. 1B). Over-
1241
1242
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
FIG. 3. Ionic dependence of APs in early fetal neurons. Current-clamp
records of voltage traces before (top) and after (bottom) application of 300 ␮M
Cd2⫹ and/or 1 ␮M TTX. Right: overlays of the membrane potential before (*)
and after drug application. A: the AP of this neuron, generated in response to
depolarizing current steps of 0.4 nA, was completely blocked after exposure to
Cd2⫹. B: the AP of another neuron, generated in response to depolarizing
current steps of 0.1 nA, was completely blocked after exposure to TTX. C: this
neuron had a RMP of ⫺40 mV but was held hyperpolarized (HYP) at around
⫺90 mV for the duration of the drug applications. Depolarizing current steps
of 0.1 nA were used for the top 2 traces, while larger steps were applied in the
bottom 2 traces (0.3 , and 0.25 nA during wash-out). Exposure to Cd2⫹ alone
reduced the amplitude of the AP (overlay shown in panel on right), although
both Cd2⫹ and TTX were needed to completely block the AP. Partial recovery
of the AP was seen after 25 min wash-out.
J Neurophysiol • VOL
tude and duration of the current step further did not elicit any
additional APs. This finding also suggests that channels underlying differences in firing properties are not expressed at these
early fetal stages of development. Six early fetal neurons
generated multiple APs from rest, while 10 others revealed a
single shunted AP only if the neuron was hyperpolarized. Since
the majority of early fetal neurons only elicited a single AP in
response to depolarization, the ionic dependence of seven
immature neurons was examined using 1 ␮M TTX (voltagedependent Na⫹ channel blocker) and/or 300 ␮M Cd2⫹ (nonspecific Ca2⫹ channel blocker). The APs of two neurons were
completely blocked with Cd2⫹ alone (Fig. 3A), another two
were blocked with TTX alone (Fig. 3B), while three required
the combined presence of TTX and Cd2⫹ (Fig. 3C). These
results suggest that the APs of some neurons are initially Na⫹
dependent, while others are initially dependent on Ca2⫹.
From mid fetal stages onward, two main types of neurons
could be identified by their firing properties in response to
depolarizing current injections, as has been previously reported
in adult guinea pig sympathetic neurons (Cassell and
McLachlan 1987; Cassell et al. 1986). Approximately onethird of mid fetal neurons (12 of 32; Fig. 5A) and one-third of
late fetal, neonatal and adult neurons (22 of 74; Fig. 7A) fired
APs throughout the duration of a depolarizing current injection, similar to adult tonic (slowly adapting) neurons (Cassell
et al. 1986; McLachlan and Meckler 1989). At all developmental stages, these tonic-firing neurons were primarily located within medial regions of the celiac ganglion, as seen in
mature guinea pigs (McLachlan and Meckler 1989).
Another third of the mid fetal neurons (11 of 32; Fig. 6A)
and more than half of the neurons at late fetal, neonatal, and
adult stages (42 of 74; Fig. 8A) adapted rapidly at the onset of
a suprathreshold depolarizing current step, as occurs in adult
phasic neurons (Cassell et al. 1986). At mid fetal and later
stages of development, 64 and 76% of the phasic neurons,
respectively, had a LAH lasting ⱖ1 s, typical of adult LAH
neurons (Cassell and McLachlan 1987). Regardless of developmental stage, the majority of phasic-firing neurons were
located in the lateral regions of the celiac ganglion as seen in
mature guinea pigs (McLachlan and Meckler 1989).
The remaining 9 of the 32 mid fetal neurons examined
resembled early fetal neurons as they had single, small amplitude APs and could not be definitively classified by their firing
properties. In contrast, only 8 of the 74 neurons from late fetal
and subsequent stages could not be classified by their firing
properties alone (see also, Cassell et al. 1986; Keast et al. 1993;
McLachlan and Meckler 1989; Stebbing and Bornstein 1993).
DEVELOPMENT OF THE AHP. In mature guinea pigs, most sympathetic neurons have a prominent AHP that is largely due to
the presence of a Ca2⫹-dependent K⫹ current, IAHP (sometimes
called gKCa1) (Cassell and McLachlan 1987; Cassell et al.
1986). In addition to this current, LAH neurons also have a
second Ca2⫹-dependent K⫹ current that is responsible for the
prolonged phase of the AHP (IsAHP or gKCa2) (Cassell and
McLachlan 1987; Jobling et al. 1993). Here we consider the
developmental appearance of each phase of the AHP. A summary of the results obtained using current-clamp recordings,
which show the combined effects of these two currents, is
shown for neurons at different stages of development in Table 2.
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
the RMP becomes more negative with age (Fig. 1A). Finally,
there was a small but significant decrease in the AP half-width
during development as reported in other developing neurons
(Fig. 2B) (Spitzer and Ribera 1998).
In adult sympathetic neurons, differences in neuronal firing
properties result from the differential expression of voltagedependent K⫹ channels and Ca2⫹-dependent K⫹ channels
(Adams and Harper 1995). Therefore the identification of
neurons with different firing properties provided the first step
in determining when the mature combinations of channels are
expressed during fetal development. We therefore investigated
the responses of developing neurons to maintained depolarizing current steps. The majority of early fetal neurons (34 of 50)
generated only a single AP with a short amplitude and long
duration in response to a depolarizing voltage step (Figs. 3 and
4A). This presumably reflects the smaller electrical driving
force driving Na⫹ and the inactivation of some Na⫹ channels
at RMP in the range ⫺30 to ⫺40 mV. Increasing the magni-
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
ranged from 28 to 148 ms (95.2 ⫾ 24.7 ms, n ⫽ 7; Fig. 5Ae).
The amplitude of the IAHP in mid fetal LAH neurons ranged
from 20 to 140 pA (73.3 ⫾ 28.3, n ⫽ 4; Fig. 6Ad), but the time
constant of this current could not be measured in these cells
due to the presence of the prolonged IsAHP.
Slow AHP (IsAHP). No evidence of slow AHPs lasting ⱖ1 s
was found in early fetal neurons. From mid fetal stages, neurons with slow AHPs characteristic of adult LAH neurons were
present. The duration of the AHP in LAH neurons decreased
by ⬃50% during development from ⬃3.5 s to ⬃2.5 s (Fig. 2D;
Table 2). The peak amplitude of IsAHP in mid fetal neurons
ranged from 20 to 60 pA (41.3 ⫾ 8.4 pA, n ⫽ 4), which was
somewhat less than that reported for LAH neurons from mature
guinea pigs (100 pA, Cassell and McLachlan 1987; 56 pA,
Martı́nez-Pinna et al. 2000). However, the decay time constant
of the underlying current in mid fetal LAH neurons (1.6 ⫾
0.1 s, n ⫽ 4; Fig. 6A) was similar to that reported for mature
guinea pigs (1.4 s, Cassell and McLachlan 1987; 1.2 s, Martı́nez-Pinna et al. 2000). This suggests that the reduction in the
duration of the AHP during development is unlikely to be due
to changes in the kinetics of IsAHP.
CURRENT-VOLTAGE RELATIONSHIPS. When early and mid fetal
neurons were injected with depolarizing or hyperpolarizing
currents small enough to alter the membrane potential by
10 –20 mV, the membrane potential often shifted back toward
rest during the current injection (Fig. 4Ab). This sag in the
membrane potential suggests the deactivation of a voltagedependent current that is active around RMP. Such a voltagedependent current is characteristic of the time-dependent rectification produced by the closure of M channels (IM) (Adams
and Harper 1995; Brown 1988; Brown and Adams 1980;
Cassell et al. 1986). When neurons were held positive to ⫺60
mV in voltage clamp, hyperpolarizing voltage steps showed
inward and outward relaxations typical of M current (Figs. 4Bc
and 5Ac). This current was deactivated below ⫺60 mV as
found in other sympathetic ganglia (Brown et al. 1982; Cassell
et al. 1986; Jobling and Gibbins 1999; Wang and McKinnon
1995). The peak amplitude of IM, measured when stepped from
FIG. 4. Electrical properties of early fetal neurons. A: current-clamp
records in a neuron from a day 34 fetus (F34) with a RMP of ⫺34 mV. Aa:
changes in membrane potential (top) during depolarizing and hyperpolarizing
steps (bottom). A short-amplitude AP is generated at the beginning of a
depolarizing step, while an AP also is generated off the anode break at the end
of the hyperpolarizing steps. Ab: a depolarizing “sag” (2) can be seen during
a small hyperpolarizing current injection. B: current-voltage relationships in a
neuron from a F34 fetus with a RMP of ⫺50 mV. Ba: current-clamp records
showing changes to the membrane potential (top) during depolarizing and
hyperpolarizing current steps (bottom). APs were generated for the duration of
a suprathreshold depolarizing step. Bb: steady-state current-voltage relationship measured at the end of the voltage step. Bc: voltage-clamp records
showing changes in current (top) during 20-mV hyperpolarizing voltage steps
(bottom) evoked from holding potentials at RMP (⫺50 mV) and at ⫺65 mV.
When held at RMP, there is an inward current relaxation during the voltage
step and an outward current relaxation following repolarization to ⫺50 mV.
These current relaxations, indicating IM, are not present at holding potentials
below ⫺60 mV. Bd: AP (top) evoked during a brief depolarizing current step
(bottom), with an AHP. Be: voltage-clamp record showing the tail-current (top)
underlying the AHP, IAHP, that followed an unclamped AP (truncated) generated by a brief suprathreshold voltage step (bottom). Bf: current-clamp record
showing a depolarizing voltage “sag” (top, 2) during a large hyperpolarizing
step (bottom) and a prolonged delay in the return of the membrane potential to
rest at the end of the hyperpolarizing step (1). Bg: voltage-clamp records
showing the currents, IH and slow IA, respectively (top), responsible for the
voltage deflections seen in Bf.
J Neurophysiol • VOL
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
Fast AHP (IAHP). The amplitude and duration of the AHP
was determined in neurons where an AP was generated in
response to a brief (10 –20 ms) suprathreshold current step. In
19 of 21 early fetal neurons, where no AP could be generated
during a depolarizing step, an AHP was observed after the AP
generated at the end of a hyperpolarizing current step (i.e., off
the “anode break”). Overall, there was a significant increase in
AHP amplitude from ⬃7 mV at mid fetal stages to ⬃15 mV at
subsequent stages (Fig. 2C; Table 2). There was a significant
increase in the duration of the AHP in tonic neurons from
⬃245 ms at mid fetal stages to ⱖ310 ms at all later stages (Fig.
2D; Table 2).
The amplitude and time constant of the current underlying
this AHP, IAHP, were measured in early fetal and mid fetal
neurons. In early fetal neurons, IAHP ranged from 17 to 45 pA
(mean 29.7 ⫾ 8.2 pA, n ⫽ 3) while the time constant ranged
from 13 to 168 ms (91.5 ⫾ 44.8 ms, n ⫽ 3; Fig. 4Be). In mid
fetal tonic neurons, the peak amplitude of IAHP ranged from 49
to 230 pA (108.0 ⫾ 23.4 pA, n ⫽ 7) while the time constant
1243
1244
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
⫺40 to ⫺60 mV, was ⬍30 pA in three early fetal neurons.
Although all mid fetal tonic neurons were observed to have a
sag in the voltage trace, the amplitude of IM was only small
compared with adults (Cassell et al. 1986; Coggan et al. 1994).
The amplitude of IM in four mid fetal tonic neurons was ⬍30
pA (Fig. 5Ac) while another was 68 pA (mean 32.8 ⫾ 9.2 pA,
n ⫽ 5). Two mid fetal LAH neurons had IM amplitudes of 16
and 74 pA.
J Neurophysiol • VOL
Half of the early fetal neurons, but only 26% of mid fetal
neurons, showed a noticeable depolarizing sag in their voltage
trace during large current injections that hyperpolarized the
neuron below ⫺100 mV (Figs. 4Bf and 5Af; Table 3). In
voltage clamp, slowly activating inward currents were observed when the holding potential was stepped below ⫺100
mV (Figs. 4Bg and 5Ag). These currents resembled IH (sometimes called IQ or If), which has previously been described in
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
FIG. 5. Electrical and morphological properties
of a mid fetal (F44) tonic neuron. Aa: currentclamp records showing changes in membrane potential (top) during depolarizing and hyperpolarizing current steps (bottom) in a tonic neuron with
RMP of ⫺43 mV. This neuron discharged APs for
the duration of a suprathreshold depolarizing current step. An A-current “notch” (1) is seen at the
break of the largest hyperpolarizing current steps.
Ab: current-voltage relationship. Ac: voltage-clamp
records showing currents (top) during 20-mV voltage steps (bottom) evoked when the neuron was
held at rest (⫺43 mV) or hyperpolarized to ⫺58
mV. Inward and outward current relaxations are
seen during the voltage step when the neuron is
held at ⫺43 mV but are abolished when held at
⫺58 mV. Ad: AP evoked after a 10-ms suprathreshold depolarizing current injection. Ae: voltageclamp record of an unclamped AP (truncated) followed by the tail-current which underlies the AHP
(IAHP). Af: current-clamp record of membrane potential (top) during and after a large hyperpolarizing current step (bottom). Note the depolarizing
voltage sag during the hyperpolarizing step (2),
and the prolonged delay in the return to RMP after
the end of the hyperpolarizing step (1). Ag: voltage-clamp record of the currents, IH and slow IA,
underlying deflections in the voltage trace seen in
Af. Ba: rendered 3D reconstruction of low-magnification confocal through-series of the same neuron
shown in A. Threshold was optimized for dendrites.
*, axon. Scale bar ⫽ 50 ␮m. Bb: rendered 3D
reconstruction of high-magnification confocal
through-series of the neuronal soma shown in Ba.
Scale bar ⫽ 10 ␮m.
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
1245
other neurons (Barrett et al. 1980; Cassell and McLachlan
1987; Cassell et al. 1986; Halliwell and Adams 1982; Jobling
and Gibbins 1999; Lüthi and McCormick 1998; Smith 1994).
Plots of current amplitude against steady-state voltage reTABLE
sponses revealed inward (or anomalous) rectification when
neurons were hyperpolarized below ⫺90 mV in about onethird of early fetal neurons (Fig. 4Bb), ⬎90% of mid fetal tonic
neurons (Fig. 5Ab, see also Fig. 7Ab), but ⬍30% of mid fetal
2. Current-clamp measurements of amplitude and duration of afterhyperpolarization (AHP) at different stages of development
AHP amplitude, mV
All neurons
Tonic neurons
LAH neurons
Phasic neurons
AHP duration, ms
All neurons
Tonic neurons
LAH neurons
Phasic neurons
Mid Fetal (F40–F45)
Late Fetal (F46⫹)
Neonatal (P0–P13)
Adult
7.5 ⫾ 0.9 (25)
10.5 ⫾ 1.5 (11)
5.9 ⫾ 0.9 (6)
6.8 ⫾ 2.0 (3)
16.1 ⫾ 3.1 (7)
16.4 ⫾ 3.4 (4)
4.0 (1)
26.4 (1)
15.0 ⫾ 0.9 (35)
15.2 ⫾ 2.5 (9)
15.1 ⫾ 1.1 (18)
14.5 ⫾ 1.6 (7)
15.4 ⫾ 1.1 (12)
15.6 ⫾ 1.4 (3)
16.3 ⫾ 1.7 (7)
14.8 (1)
1201.6 ⫾ 337.2 (22)
244.6 ⫾ 35.0 (11)
3695.0 ⫾ 187.0 (6)
295.0 ⫾ 50.6 (3)
281.5 ⫾ 42.7 (6)
313.0 ⫾ 28.4 (4)
—
90 (1)
1495.5 ⫾ 207.4 (34)
488.1 ⫾ 65.0 (9)
2471.9 ⫾ 192.0 (18)
279.9 ⫾ 33.9 (7)
1807.7 ⫾ 407.1 (12)
348.0 ⫾ 75.8 (3)
2885.7 ⫾ 230.2 (7)
410 (1)
Values are means ⫾ SE. Parentheses enclose number of neurons. LAH, long AHP.
J Neurophysiol • VOL
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
FIG. 6. Electrical and morphological properties of
a mid fetal (F44) LAH neuron. Aa: current-clamp
records showing changes in membrane potential (top)
in response to depolarizing and hyperpolarizing current steps (bottom). This neuron only discharged APs
at the onset of suprathreshold depolarizing steps.
Ab: current-voltage relationship. Ac: current-clamp
record of AP and the LAH (top), evoked after a brief
suprathreshold depolarizing current step (bottom).
Ad: voltage-clamp record of an unclamped action
current and tail currents underlying the AHP (top,
filtered at 450 Hz), generated after a brief (10 ms)
depolarizing voltage step (bottom). Ba: rendered 3D
reconstruction of low-magnification confocal throughseries of the same neuron whose electrical properties
are shown in A. Threshold was optimized for dendrites. *, axon. Scale bar ⫽ 20 ␮m. Bb: rendered 3D
reconstruction of high-magnification confocal throughseries of the neuronal soma.
1246
TABLE
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
greater in medial regions (P ⬍ 0.05) so that by adult stages, the
relative area occupied by neuropil in medial regions (50 ⫾ 2%;
n ⫽ 5 animals) was significantly higher than that in lateral
regions (33.6 ⫾ 4.1%; n ⫽ 5 animals, P ⬍ 0.05; Fig. 9A).
3. Proportion of neurons with proposed ionic currents
Early Fetal
(F30–F34,*
1–7 g)
IM
IH
IA
Slow IA
IKi
Mid Fetal (F38–F44,* 10–49 g)
Immature†
Immature†
Tonic
LAH
Phasic
15/21 (71)
9/17 (53)
10/19 (53)
9/22 (41)
5/17 (29)
7/11 (64)
1/5 (20)
4/8 (50)
0/7 (0)
0/5 (0)
15/15 (100)
3/10 (30)
12/15 (80)
6/15 (40)
11/12 (92)
5/7 (71)
1/6 (17)
4/7 (57)
0/7 (0)
2/7 (29)
3/4 (75)
1/2 (50)
2/3 (67)
0/3 (0)
0/3 (0)
Values in parentheses are percentages. * Gestational age range for animals
used in this part of the study. † Could not be classified according to firing
properties.
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
LAH neurons (Fig. 6Ab, see also Fig. 8Ab; Table 3). The
voltage-dependent K⫹ current, IKi, responsible for this rectification has been identified in many autonomic neurons where it
is predominantly restricted to tonic-firing neurons (Adams and
Harper 1995; Cassell and McLachlan 1987; Cassell et al. 1986;
Keast et al. 1993; Wang and McKinnon 1995).
After neurons were hyperpolarized to potentials below ⫺60
mV, a prolonged delay was often seen in the voltage trace as
the membrane returned to rest (Fig. 5Aa). This delay or “notch”
has been described in mature sympathetic neurons, where it is
due to the activation of a transient outward A current (IA)
(Adams and Galvan 1986; Cassell et al. 1986; Connor and
Stevens 1971; Galvan and Sedlmeir 1984; Wang and McKinnon 1995). Around half of the early fetal neurons and 80% of
mid fetal tonic neurons appeared to express IA (Table 3). In 10
mid fetal tonic neurons that were voltage clamped at rest, a
transient outward IA was seen when the holding potential
returned to rest (around ⫺50 mV) after being held below ⫺60
mV. The time constant of IA inactivation in these neurons
ranged from 7 to 20 ms (12.9 ⫾ 2.1 ms, n ⫽ 6), which is less
than that reported for mature guinea-pig sympathetic tonic
neurons (mean 22.1 ms, n ⫽ 17) (Cassell et al. 1986).
An outward current, slower and more prolonged than IA
(slow IA) was observed in 41% of early fetal neurons, including some with single shunted action potentials (Fig. 4B, f and
g), and 40% of mid fetal tonic neurons (Fig. 5A, f and g; Table
3). This current was not seen in LAH or phasic neurons at mid
fetal stages, reflecting the situation in mature guinea-pigs (Table 3) (Cassell et al. 1986). Similar currents have been observed in developing rat sympathetic neurons (IAs) (McFarlane
and Cooper 1992) where, in mature animals, its expression also
is restricted to tonic neurons (ID2) (Wang and McKinnon
1995).
Development of neuronal morphology
We used the
relative area of the celiac ganglion occupied by neuropil as an
indicator of dendritic growth in our stereological analysis of
the medial and lateral regions at different stages of development (Fig. 9A). At late embryonic stages, very little neuropil
was observed in either medial (area of neuropil: 4 ⫾ 4% of
total sample area, n ⫽ 3 embryos) or lateral regions (6 ⫾ 4%;
n ⫽ 3 embryos). The relative area occupied by neuropil significantly increased throughout development in both medial
and lateral regions. However, the rate of increase observed was
PROPORTION OF AREA OCCUPIED BY NEUROPIL.
J Neurophysiol • VOL
FIG. 7. Electrical and morphological properties of a late fetal (F49) tonic
neuron. Aa: current-clamp record showing continuous discharge of APs (top)
in response to a suprathreshold depolarizing current step (bottom). Ab: currentvoltage relationship showing marked inward rectification below ⫺90 mV. Ac:
a 10-ms depolarizing current step elicited a single AP that was followed by an
AHP with a duration ⬍500 ms. Trace filtered at 150 Hz. B: rendered 3D
reconstruction of a low-magnification confocal through-series of the neuron.
Threshold was optimized for dendrites. *, axon. Scale bar ⫽ 50 ␮m.
86 • SEPTEMBER 2001 •
www.jn.org
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
1247
ELECTROPHYSIOLOGICAL CLASS, MORPHOLOGY, AND NEUROPEPTIDE CONTENT. Using a subset of neurons described above,
FIG. 8. Electrical and morphological properties of a neonatal (P2) LAH
neuron. Aa: current-clamp record showing brief discharge of APs (top) at the
onset of a suprathreshold depolarizing current step (bottom). Ab: currentvoltage relationship. Ac: a brief depolarizing current step elicited a single AP
that was followed by a prolonged AHP that lasted several seconds. Trace
filtered at 150 Hz. B: rendered 3D reconstruction of a low-magnification
confocal through-series of the neuron. Threshold was optimized for dendrites.
*, axon. Scale bar ⫽ 50 ␮m.
To gain a more detailed
analysis of neuronal morphology, individual Neurobiotin-filled
neurons, some of whose electrical properties had been anaOVERALL NEURONAL MORPHOLOGY.
J Neurophysiol • VOL
differences between the morphology of dye-filled tonic and
LAH neurons were examined from mid fetal through to neonatal stages. As previously described in mature guinea pigs
(Gibbins et al. 1999; Keast et al. 1993), tonic neurons were
located in medial regions while LAH neurons were located in
lateral regions (␹2 ⫽ 5.3, df ⫽ 1, P ⫽ 0.02, n ⫽ 53 neurons).
While no differences were found in the cross-sectional areas of
neuronal cell bodies (tonic, 844.5 ⫾ 103.2 ␮m2, n ⫽ 16; LAH,
1,159.2 ⫾ 100.0 ␮m2, n ⫽ 18; F(1,31) ⫽ 0.91, P ⫽ 0.4), tonic
neurons had more primary dendrites (12.9 ⫾ 1.0, n ⫽ 12)
compared with LAH neurons (7.3 ⫾ 0.6, n ⫽ 16; F(1,25) ⫽
19.7, P ⬍ 0.001) as well as greater total dendritic lengths
(tonic, 2,299.3 ⫾ 345.6 ␮m, n ⫽ 12; cf. LAH, 848.2 ⫾ 65.6
␮m, n ⫽ 16; F(1,25) ⫽ 23.9, P ⬍ 0.001; Fig. 9D). Consequently, the soma of tonic neurons formed a significantly
smaller proportion of the total neuronal surface area (6.5 ⫾
0.9%, n ⫽ 9) compared with LAH neurons (16.6 ⫾ 2.2%, n ⫽
12; t-test, df ⫽ 19, P ⫽ 0.001).
The neuropeptide content of 101 dye-filled neurons at fetal
and neonatal stages was determined. At mid fetal stages, 72%
of neurons without NPY-IR, with or without Som-IR, were
located in medial regions while the remaining neurons were
located in lateral regions (n ⫽ 32). Across all stages examined,
only 1 of 17 tonic neurons expressed NPY-IR, while 6 of 10
LAH neurons contained NPY-IR. Combined analysis of neurons from fetal and neonatal stages revealed that the cell body
cross-sectional area of neurons without NPY-IR, many of
which contained Som-IR, was only marginally greater than
neurons with NPY-IR (Mann-Whitney U ⫽ 371.5, P ⫽ 0.05,
n ⫽ 94). The number of primary dendrites on neurons with and
without NPY-IR was not significantly different (Mann-Whit-
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
lyzed (Figs. 5B, 6B, 7B, and 8B), were examined. It was
difficult to achieve reliable fills of early fetal neurons, presumably due to the short impalement times. Therefore the analysis
of dendritic fields is largely restricted to mid fetal and subsequent stages. There was a small but significant increase in the
number of primary dendrites (Fig. 9B) and total dendritic
length (Fig. 9C) during these stages. Consistent with the stereological analyses, by late fetal stages, medially located neurons had more primary dendrites (12.4 ⫾ 1.4, n ⫽ 7) and
greater total dendritic lengths (1,681 ⫾ 330.0 ␮m, n ⫽ 7)
compared with laterally located neurons with 6.6 ⫾ 1.7 (n ⫽ 5)
primary dendrites and total dendritic lengths of 674.4 ⫾ 246.1
␮m (n ⫽ 5).
There was a dramatic increase in the cross-sectional area of
neuronal cell bodies during development from 170 ␮m2 at
early fetal stages to ⬎1,000 ␮m2 at neonatal stages (Fig. 9D).
The cell body cross-sectional area of neurons was similar
regardless of their topographical location within the ganglion
(Mann-Whitney U test ⫽ 2,386.5, P ⫽ 0.7, n ⫽ 147). When
the morphological and electrical properties of individual neurons were determined, the surface area and capacitance were
calculated. Overall there was no significant correlation between the total surface area (neuronal soma and dendritic field)
of a neuron and the derived input capacitance (Fig. 9E). However, the increasing surface area of neuronal soma was strongly
correlated with an increase in the derived input capacitance
(Fig. 9F).
1248
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
ney U ⫽ 78.0, P ⫽ 0.08, n ⫽ 53). In contrast, neurons without
NPY-IR, including those with Som-IR, had significantly
greater total dendritic lengths than those with NPY-IR (MannWhitney U ⫽ 24.0, P ⫽ 0.003, n ⫽ 42; Fig. 9C).
Overall the correlations between electrophysiological,
morphological, and neurochemical properties of fetal neu-
rons reflect those previously published for celiac ganglion
neurons from mature guinea pigs. Thus NPY-IR neurons
corresponded to LAH neurons with small dendritic fields,
while Som-IR neurons corresponded to tonic neurons with
large dendritic fields. Nevertheless neonatal neurons of all
classes were still only about two-thirds the size of neurons
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
J Neurophysiol • VOL
86 • SEPTEMBER 2001 •
www.jn.org
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
in adult celiac ganglia (Boyd et al. 1996; Gibbins et al.
1999; Keast et al. 1993).
DISCUSSION
We have shown that different functional subpopulations of
celiac ganglion neurons can be distinguished by their electrical
and morphological properties from mid fetal stages of development. These distinctions occur after the neurochemical phenotypes of the neurons and the topographical organization of
the celiac ganglion have been established (Anderson et al.
2001) but long before the neurons finish growing. The differentiation of these neurons involves the sequential expression of
various K⫹ channels accompanied by divergent growth patterns of their dendritic trees. Furthermore each major functional class of neuron seems to develop directly from a topographically distinct subset of precursors.
istic patterns of expression of K⫹ channels. Most notably, M
current is largely restricted to phasic/LAH neurons while A
current regulates AP discharge in tonic neurons but not LAH
neurons (Cassell and McLachlan 1987; Cassell et al. 1986;
Wang and McKinnon 1995). In contrast with adult neurons,
small M currents were detected in most early fetal neurons and
in both phasic and tonic firing neurons at mid fetal stages.
Consequently there must be a significant increase in M-current
expression in phasic/LAH neurons but not tonic neurons during
the later stages of fetal development. In mature sympathetic
neurons, M current is thought to exert a major influence on
firing properties by reducing the rate of action potential generation (Adams and Harper 1995; Wang and McKinnon 1995).
The low level of expression of M current in developing celiac
ganglion neurons suggests that it has only a limited influence
on their firing properties. At early fetal stages, 40% of celiac
ganglion neurons expressed the slow A current. By mid fetal
stages, the slow A current was restricted to tonic neurons,
suggesting that the early expression of this current provides the
first indication that a neuron is destined to develop the tonicfiring phenotype.
In contrast to the early expression of M and A currents, the
IsAHP responsible for the LAH was not detected until mid fetal
stages. This explains the relatively late stage at which LAH
neurons could be identified by functional criteria. The late
expression of Ca2⫹-dependent K⫹ channels also has been
reported in other systems (Ahmed et al. 1986; Dryer 1994,
1998; Martin-Caraballo and Greer 2000). It has been suggested
previously that ion channels required for basic neuronal excit-
ability (such as voltage-dependent Na⫹, Ca2⫹, and K⫹ channels) are established relatively early during development and
are not influenced by external factors (Dryer 1994; Ribera and
Spitzer 1992). However, the developmental expression of ion
channels involved in the fine control of neuronal firing behavior (such as the Ca2⫹-dependent K⫹ channel IsAHP, as well as
IA and IM) are likely to be influenced by extrinsic factors
including the local environment, synaptic inputs, and targets
(Barish 1995; Dryer 1994, 1998; McFarlane and Cooper 1992;
Raucher and Dryer 1994, 1995). If so, the sequential expression of ion channels during development of celiac ganglion
neurons implies the presence of multiple factors acting in a
time-dependent way to regulate their differentiation.
ELECTROPHYSIOLOGICAL AND MORPHOLOGICAL PHENOTYPES
ARE ESTABLISHED DURING THE SAME DEVELOPMENTAL PERIOD.
The celiac ganglion neurons developed their characteristic
electrical and morphological phenotypes in parallel, mainly
during the mid fetal period. Such parallel development has
been reported widely in other neurons (Allan and Greer
1997a,b; Dekkers et al. 1994; Kandler and Friauf 1995; MartinCaraballo and Greer 1999; Phelan et al. 1997; Vincent and Tell
1999; Warren and Jones 1997). However, the electrical properties of celiac ganglion neurons did not change after they were
established at mid fetal stages, whereas neurons continued to
increase in size. Therefore as the neurons grow during late fetal
and neonatal development, there must be continued regulated
synthesis of phenotypically appropriate channels to match the
on-going production of new cell membrane.
Much of the growth of the neurons involves the dendritic
tree as well as the soma. Nevertheless developmental increases
in apparent cell capacitance were correlated much more
strongly with somatic surface area rather than with the total
surface area of the neurons including their dendrites. The
simplest interpretation of this observation is that most of the
electrical properties we recorded arose from the somatic membrane with relatively little contribution from the dendrites. This
is surprising since it is generally thought that mature autonomic
neurons, which are significantly larger than the fetal neurons,
are electrotonically compact (Adams and Harper 1995).
While celiac ganglion neurons from medial and lateral regions showed similar developmental increases in somatic size,
the dendritic trees of neurons located in medial regions of the
ganglion increased their total length at a greater rate than those
in lateral regions. This differential growth results in tonic
neurons in the medial regions of the ganglion bearing larger
dendritic trees than LAH neurons in the lateral regions of the
ganglion. The dissociation of somatic and dendritic growth
FIG. 9. Developmental changes in morphological properties. A: relationship during development between relative area occupied
by neuropil in lateral and medial regions of the celiac ganglion, expressed as a percentage of the total sample area. While there was
a significant increase in the relative area occupied by neuropil in both lateral (R2 ⫽ 0.54, F(1,23) ⫽ 27.0, P ⬍ 0.0001) and medial
(R2 ⫽ 0.73, F(1,24) ⫽ 63.8, P ⬍ 0.0001) regions, the rate of increase was significantly greater in medial regions (*P ⬍ 0.05). Stages
of development are indicated under corresponding weights (E, embryonic; F, fetal; Neo, neonate). B: there was a small but
significant increase in the number of primary dendrites (ⱖ1 cell body diameter in length) during development (R2 ⫽ 0.06, F(1,79) ⫽
4.9, P ⫽ 0.03). C: a small but significant increase in the total dendritic length was seen during development (R2 ⫽ 0.08, F(1,58) ⫽
4.9, P ⫽ 0.03). Symbols indicate neurons whose electrical firing properties and neurochemical content also were determined. Note
that LAH neurons had smaller total dendritic lengths than tonic neurons. Also note, tonic neurons were all NPY⫺ve, compared with
LAH neurons, which were NPY⫹ve or NPY⫺ve. D: there was a significant increase in the cross-sectional (XS) area of neuronal
cell bodies during development (R2 ⫽ 0.66, F(1,168) ⫽ 329.8, P ⬍ 0.001). E: there was no significant correlation between the derived
capacitance (pF) and the total surface area (␮m2; neuronal soma and dendritic field) of neurons (R2 ⫽ 0.04, F(1,24) ⫽ 0.92, P ⫽
0.3). F: there was a significant correlation between the derived capacitance (pF) and the surface area of the neuronal soma (␮m2;
R2 ⫽ 0.49, F(1,13) ⫽ 12.7, P ⫽ 0.004). Neurons with higher capacitance measurements had soma with greater surface areas.
J Neurophysiol • VOL
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
EARLY FETAL NEURONS EXPRESS DIFFERENT COMBINATIONS OF
K⫹ CHANNELS. Adult celiac ganglion neurons have character-
1249
1250
R. L. ANDERSON, P. JOBLING, AND I. L. GIBBINS
We are grateful to Professor W. W. Blessing and Dr. J. Oliver for the gift of
antiserum to NPY, and to Dr. J. C. Brown (Medical Research Council of
Canada, Regulatory Peptide Group, Vancouver, British Columbia, Canada) for
the provision of antiserum to somatostatin. We also thank Dr. G. Hennig for
the use of National Institutes of Health image macros. Finally, we thank Assoc.
Prof. J. L. Morris and Dr. S.J.H. Brookes for comments on the manuscript. P.
Jobling was a National Health and Medical Research Council (NHMRC)
Australian Postdoctoral Fellow. R. L. Anderson was a recipient of an NHMRC
Dora Lush Biomedical Postgraduate Research Scholarship.
This work was supported by grants from the NHMRC (Grants 970033 and
977400), the Clive and Vera Ramaciotti Foundation, the Charles and Sylvia
Viertel Charitable Foundation, Flinders Medical Center Foundation, and the
Flinders University Research Budget.
REFERENCES
ADAMS DJ AND HARPER AA. Electrophysiological properties of autonomic
ganglion neurons. In: Autonomic Ganglia, edited by McLachlan EM. Luxembourg: Harwood, 1995, p. 153–212.
ADAMS PR AND GALVAN M. Voltage-dependent currents of vertebrate neurons
and their role in membrane excitability. Adv Neurol 44: 137–170, 1986.
AHMED Z, CONNOR JA, TANK DW, AND FELLOWS RE. Expression of membrane
currents in rat diencephalic neurons in serum-free culture. Brain Res 393:
221–231, 1986.
ALLAN DW AND GREER JJ. Embryogenesis of the phrenic nerve and diaphragm
in the fetal rat. J Comp Neurol 382: 459 – 468, 1997a.
J Neurophysiol • VOL
ALLAN DW AND GREER JJ. Development of phrenic motoneuron morphology
in the fetal rat. J Comp Neurol 382: 469 – 479, 1997b.
ANDERSON RL, MORRIS JL, AND GIBBINS IL. Neurochemical differentiation of
functionally distinct populations of autonomic neurons. J Comp Neurol 429:
419 – 435, 2001.
ANDREWS TJ, THRASIVOULOU C, NESBIT W, AND COWEN T. Target-specific
differences in the dendritic morphology and neuropeptide content of neurons
in the rat SCG during development and aging. J Comp Neurol 368: 33– 44,
1996.
BARISH ME. Modulation of the electrical differentiation of neurons by interactions with glia and other non-neuronal cells. Perspect Dev Neurobiol 2:
357–370, 1995.
BARRETT EF, BARRETT JN, AND CRILL WE. Voltage-sensitive outward currents
in cat motoneurones. J Physiol (Lond) 304: 251–276, 1980.
BOYD HD, MCLACHLAN EM, KEAST JR, AND INOKUCHI H. Three electrophysiological classes of guinea pig sympathetic postganglionic neurone have
distinct morphologies. J Comp Neurol 369: 372–387, 1996.
BROWN DA. M currents. In: Ion Channels, edited by Narahashi T. New York:
Plenum, 1988, p. 55–94.
BROWN DA AND ADAMS PR. Muscarinic suppression of a novel voltagesensitive K⫹ current in a vertebrate neurone. Nature 283: 673– 676, 1980.
BROWN DA, ADAMS PR, AND CONSTANTI A. Voltage-sensitive K-currents in
sympathetic neurons and their modulation by neurotransmitters. J Auton
Nerv Syst 6: 23–35, 1982.
BUTLER H AND JUURLINK BHJ. An Atlas for Staging Mammalian and Chick
Embryos. Boca Raton, FL: CRC, 1987.
CAMERON JS AND DRYER SE. BK-type KCa channels in two parasympathetic
cell types: differences in kinetic properties and developmental expression.
J Neurophysiol 84: 2767–2776, 2000.
CASSELL JF, CLARK AL, AND MCLACHLAN EM. Characteristics of phasic and
tonic sympathetic ganglion cells of the guinea-pig. J Physiol (Lond) 372:
457– 483, 1986.
CASSELL JF AND MCLACHLAN EM. Two calcium-activated potassium conductances in a subpopulation of coeliac neurones of guinea-pig and rabbit.
J Physiol (Lond) 394: 331–349, 1987.
CHIBA T AND TANAKA K. A target specific pathway from nitric oxide synthase
immunoreactive preganglionic sympathetic to superior cervical ganglion
neurons innervating the submandibular salivary gland. J Auton Nerv Syst 71:
139 –147, 1998.
COGGAN JS, PURNYN SL, KNOPER SR, AND KREULEN DL. Muscarinic inhibition
of two potassium currents in guinea-pig prevertebral neurons: differentiation
by extracellular cesium. Neuroscience 59: 349 –361, 1994.
CONNOR JA AND STEVENS CF. Voltage clamp studies of a transient outward
membrane current in gastropod neural somata. J Physiol (Lond) 213: 21–30,
1971.
COSTA M AND FURNESS JB. Somatostatin is present in a subpopulation of
noradrenergic nerve fibres supplying the intestine. Neuroscience 13: 911–
919, 1984.
DAVIES PJ, IRELAND DR, MARTÍNEZ-PINNA J, AND MCLACHLAN EM. Electrophysiological roles of L-type channels in different classes of guinea pig
sympathetic neuron. J Neurophysiol 82: 818 – 828, 1999.
DEKKERS J, BECKER DL, COOK JE, AND NAVARRETE R. Early postnatal changes
in the somatodendritic morphology of ankle flexor motoneurons in the rat.
Eur J Neurosci 6: 87–97, 1994.
DRYER SE. Functional development of the parasympathetic neurons of the
avian ciliary ganglion: a classic model system for the study of neuronal
differentiation and development. Prog Neurobiol 43: 281–322, 1994.
DRYER SE. Role of cell-cell interactions in the developmental regulation of
Ca2⫹-activated K⫹ currents in vertebrate neurons. J Neurobiol 37: 23–36,
1998.
DRYER SE AND CHIAPPINELLI VA. An intracellular study of synaptic transmission and dendritic morphology in sympathetic neurons of the chick embryo.
Brain Res 354: 99 –111, 1985.
GALVAN M AND SEDLMEIR C. Outward currents in voltage-clamped rat sympathetic neurones. J Physiol (Lond) 356: 115–133, 1984.
GIBBINS IL. Peripheral autonomic nervous system. In: The Human Nervous
System, edited by Paxinos G. New York: Academic, 1990, p. 93–123.
GIBBINS IL. Chemical neuroanatomy of sympathetic ganglia. In: Autonomic
Ganglia, edited by McLachlan EM. Luxembourg: Harwood, 1995, p. 73–
120.
GIBBINS IL, MATTHEW SE, AND JOBLING P. Pathway-specific expression of
PKC and PKA in sympathetic neurons. Neuroreport 10: 975–979, 1999.
HALLIWELL JV AND ADAMS PR. Voltage-clamp analysis of muscarinic excitation in hippocampal neurons. Brain Res 250: 71–92, 1982.
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
rates has been reported previously in motor neurons of postnatal rats (Nunez-Abades and Cameron 1995) and suggests that
somatic size and dendritic size may be regulated independently
during development.
The medially located tonic-firing neurons in the mature
guinea pig celiac ganglion are unusual in that they receive
convergent synaptic inputs from neurons projecting from the
gut wall (enteric intestinofugal neurons) in addition to synaptic
inputs from preganglionic neurons projecting from the spinal
cord (Kreulen and Szurszewski 1979; McLachlan and Meckler
1989; Meckler and McLachlan 1988). Previous studies on
developing autonomic neurons have demonstrated a close temporal relationship between dendritic outgrowth and synapse
formation (Dryer 1994; Hirst and McLachlan 1986; Rubin
1985a– c). Therefore we predict that synaptic inputs are established during the same developmental period in which differential growth of the dendrites is occurring, i.e., from early to
mid fetal stages of development, and that the increased dendritic growth rate of the medial neurons is related to the arrival
of additional peripheral inputs, which the smaller lateral neurons lack. If this prediction is borne out, the initiation of the
differential expression of at least some K⫹ channels, such as
those responsible for the A current, may well precede synaptogenesis in the celiac ganglion.
In conclusion, we have shown that the differentiation of
electrical and morphological properties of sympathetic neurons
in the celiac ganglion follows the development of their neurochemical phenotype. The selective expression of K⫹ channels
follows a step-wise sequence that occurs simultaneously with
the differential growth of the dendritic trees of specific populations of neurons innervating the vasculature or the enteric
plexuses. This sequential development combined with the dissociation between somatic and dendritic growth of the neurons
strongly implies that many of these phenotypic traits can be
independently regulated. Thus our data provide strong circumstantial evidence that the development of the full phenotype of
different functional classes of autonomic final motor neurons is
a multi-step process, likely to involve a regulated sequence of
trophic interactions.
DEVELOPMENT OF SUBPOPULATIONS OF AUTONOMIC NEURONS
J Neurophysiol • VOL
MORRIS JL, KONDO M, AND GIBBINS IL. Selective innervation of different target
tissues in guinea-pig cranial exocrine glands by sub-populations of parasympathetic and sympathetic neurons. J Auton Nerv Syst 66: 75– 86, 1997.
MORRIS JL, ZHU BS, GIBBINS IL, AND BLESSING WW. Subpopulations of
sympathetic neurons project to specific vascular targets in the pinna of the
rabbit ear. J Comp Neurol 412: 147–160, 1999.
NERBONNE JM AND GURNEY AM. Development of excitable membrane properties in mammalian sympathetic neurons. J Neurosci 9: 3272–3286, 1989.
NUNEZ-ABADES PA AND CAMERON WE. Morphology of developing rat genioglossal motoneurons studied in vitro: relative changes in diameter and
surface area of somata and dendrites. J Comp Neurol 353: 129 –142, 1995.
PHELAN KD, CHANG JY, AND KANE CJ. Transforming growth factor-beta2
selectively alters the developmental expression of the fast transient A-current in cultured rat superior cervical ganglion neurons. J Neurosci Res 49:
475– 484, 1997.
RAUCHER S AND DRYER SE. Functional expression of A-currents in embryonic
chick sympathetic neurones during development in situ and in vitro.
J Physiol (Lond) 479: 77–93, 1994.
RAUCHER S AND DRYER SE. Target-derived factors regulate the expression of
Ca2⫹-activated K⫹ currents in developing chick sympathetic neurones.
J Physiol (Lond) 486: 605– 614, 1995.
RIBERA AB AND SPITZER NC. Developmental regulation of potassium channels
and the impact on neuronal differentiation. In: Ion Channels, edited by
Narahashi T. New York: Plenum, 1992, p. 1–38.
RUBIN E. Development of the rat superior cervical ganglion: ganglion cell
maturation. J Neurosci 5: 673– 684, 1985a.
RUBIN E. Development of the rat superior cervical ganglion: ingrowth of
preganglionic axons. J Neurosci 5: 685– 696, 1985b.
RUBIN E. Development of the rat superior cervical ganglion: initial stages of
synapse formation. J Neurosci 5: 697–704, 1985c.
SCOTT JP. The embryology of the guinea pig. I. A table of normal development. Am J Anat 60: 397– 432, 1937.
SMITH PA. Amphibian sympathetic ganglia: an owner’s and operator’s manual.
Prog Neurobiol 43: 439 –510, 1994.
SPITZER NC. A developmental handshake: neuronal control of ionic currents
and their control of neuronal differentiation. J Neurobiol 22: 659 – 673,
1991.
SPITZER NC AND RIBERA AB. Development of electrical excitability in embryonic neurons: mechanisms and roles. J Neurobiol 37: 190 –197, 1998.
STEBBING MJ AND BORNSTEIN JC. Electrophysiological analysis of the convergence of peripheral inputs onto neurons of the coeliac ganglion in the guinea
pig. J Auton Nerv Syst 46: 93–105, 1993.
STOCKARD CR AND PAPANICOLOU GN. The existence of a typical oestrous cycle
in the guinea-pig—with a study of its histological and physiological
changes. Am J Anat 22: 225–283, 1917.
STOFER WB AND HORN JP. Expression of neuropeptide-Y-like immunoreactivity
begins after adrenergic differentiation and ganglionic synaptogenesis in
developing bullfrog sympathetic neurons. J Neurosci 10: 3305–3312, 1990.
STOFER WD AND HORN JP. Neurogenesis and differentiation of sympathetic B
and C cells in the bullfrog tadpole. J Neurosci 13: 801– 807, 1993.
VINCENT A AND TELL F. Postnatal development of rat nucleus tractus solitarius
neurons: morphological and electrophysiological evidence. Neuroscience
93: 293–305, 1999.
WANG HS AND MCKINNON D. Potassium currents in rat prevertebral and
paravertebral sympathetic neurones: control of firing properties. J Physiol
(Lond) 485: 319 –335, 1995.
WARREN RA AND JONES EG. Maturation of neuronal form and function in a
mouse thalamo-cortical circuit. J Neurosci 17: 277–295, 1997.
WEIR BJ. Reproductive characteristics of hystricomorph rodents. Symp Zool
Soc (Lond) 34: 265–301, 1974.
86 • SEPTEMBER 2001 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
HILLE B. Ionic Channels of Excitable Membranes (2nd ed.). Sunderland, MA:
Sinauer, 1992.
HIRST GDS AND MCLACHLAN EM. Post-natal development of ganglia in the
lower lumbar sympathetic chain of the rat. J Physiol (Lond) 349: 119 –134,
1984.
HIRST GDS AND MCLACHLAN EM. Development of dendritic calcium currents
in ganglion cells of the rat lower lumbar sympathetic chain. J Physiol (Lond)
377: 349 –368, 1986.
HOWARD CV AND REED BMG. Unbiased Stereology: Three-Dimensional Measurement in Microscopy. Oxford, UK: BIOS Scientific Publishers, 1998.
JÄNIG W AND MCLACHLAN EM. Specialized functional pathways are the
building blocks of the autonomic nervous system. J Auton Nerv Syst 41:
3–14, 1992.
JOBLING P AND GIBBINS IL. Electrophysiological and morphological diversity
of mouse sympathetic neurons. J Neurophysiol 82: 2747–2764, 1999.
JOBLING P, MCLACHLAN EM, AND SAH P. Calcium induced calcium release is
involved in the afterhyperpolarization in one class of guinea pig sympathetic
neurone. J Auton Nerv Syst 42: 251–257, 1993.
KANDLER K AND FRIAUF E. Development of electrical membrane properties
and discharge characteristics of superior olivary complex neurons in fetal
and postnatal rats. Eur J Neurosci 7: 1773–1790, 1995.
KEAST JR, MCLACHLAN EM, AND MECKLER RL. Relation between electophysiological class and neuropeptide content of guinea pig sympathetic prevertebral neurons. J Neurophysiol 69: 384 –394, 1993.
KREULEN DL AND SZURSZEWSKI JH. Nerve pathways in celiac plexus of the
guinea pig. Am J Physiol Endocrinol Metab 237: E90 –E97, 1979.
LINDH B, HÖKFELT T, ELFVIN L-G, TERENIUS L, FRANKENKRUG J, ELDE R, AND
GOLDSTEIN M. Topography of NPY-, somatostatin-, and VIP-immunoreactive, neuronal subpopulations in the guinea pig celiac-superior mesenteric
ganglion and their projection to the pylorus. J Neurosci 6: 2371–2383, 1986.
LÜTHI A AND MCCORMICK DA. H-current: properties of a neuronal and
network pacemaker. Neuron 21: 9 –12, 1998.
MACRAE IM, FURNESS JB, AND COSTA M. Distribution of subgroups of noradrenaline neurons in the coeliac ganglion of the guinea-pig. Cell Tissue Res
244: 173–180, 1986.
MARTIN-CARABALLO M AND GREER JJ. Electrophysiological properties of rat
phrenic motoneurons during perinatal development. J Neurophysiol 81:
1365–1378, 1999.
MARTIN-CARABALLO M AND GREER JJ. Development of potassium conductances in perinatal rat phrenic motoneurons. J Neurophysiol 83: 3497–3508,
2000.
MARTÍNEZ-PINNA J, DAVIES PJ, AND MCLACHLAN EM. Diversity of channels
involved in Ca2⫹ activation of K⫹ channels during the prolonged AHP in
guinea pig sympathetic neurons. J Neurophysiol 84: 1346 –1354, 2000.
MATSUMOTO SG, GRUENER RP, AND KREULEN DL. Neurotransmitter properties
of guinea-pig sympathetic neurons grown in dissociated cell culture. II. Fetal
and embryonic neurons: regulation of neuropeptide Y expression. Neuroscience 57: 1147–1157, 1993.
MCFARLANE S AND COOPER E. Postnatal development of voltage-gated K
currents on rat sympathetic neurons. J Neurophysiol 67: 1291–1300, 1992.
MCFARLANE S AND COOPER E. Extrinsic factors influence the expression of
voltage-gated K currents on neonatal rat sympathetic neurons. J Neurosci
13: 2591–2600, 1993.
MCLACHLAN E AND MECKLER R. Characteristics of synaptic input to three
classes of sympathetic neurone in the coeliac ganglion of the guinea-pig.
J Physiol (Lond) 415: 109 –129, 1989.
MECKLER RL AND MCLACHLAN EM. Axons of peripheral origin preferentially
synapse with tonic neurones in the guinea pig coeliac ganglion. Neuroscience Lett 86: 189 –194, 1988.
MORRIS JL, GRASBY DJ, ANDERSON RL, AND GIBBINS IL. Neurochemical
distinction between skeletal muscle vasodilator neurons and pelvic vasodilator neurons in guinea-pigs. J Auton Nerv Syst 71: 64 – 68, 1998.
1251