THE JOURNAL OF COMPARATIVE NEUROLOGY 373~1-10 (1996)
Initial Formation and Secondary
Condensation of Nerve Pathways
in the Medicinal Leech
JOHN JELLIES, DIANE M. KOPP, KRISTEN M. JOHANSEN,
AND JORGEN JOHANSEN
Department of Biological Sciences, Western Michigan University, Kalamazoo,
Michigan 49008 (J.Je.1; Neurobiology Research Center and Department of Physiology and
Biophysics, University of Alabama at Birmingham, Birmingham, Alabama 35294 (D.M.K.);
Department of Zoology and Genetics, Iowa State University, Ames, Iowa 50011 (K.M.J., J.Jo.1
ABSTRACT
Invertebrates have proved to be important experimental systems for examining questions
related to growth cone navigation and nerve formation, in large part because of their simpler
nervous systems. However, such apparent simplicity can be deceiving because the final
stereotyped patterns may be the result of multiple developmental mechanisms and not
necessarily the sole consequence of the pathway choices of individual growth cones. We have
examined the normal sequence of events that are involved in the formation of the major
peripheral nerves in leech embryos by employing (1) an antibody directed against acetylated
tubulin to label neurons growing out from the central nervous system, (2) the Lan3-2 antibody
to label a specific population of peripheral neurons growing into the central nervous system,
and (3) intracellular dye filling of single cells. We found that the mature pattern of nerves was
characterized by a pair of large nerve roots, each of which branched into two major tracts. The
earliest axonal projections did not, however, establish this pattern definitively. Rather, each of
the four nerves initially formed as discrete, roughly parallel tracts without bifurcation, with the
final branching pattern of the nerve roots being generated by a secondary condensation. In
addition, we found that some of the nerves were pioneered in different ways and by different
groups of neurons. One of the nerves was established by central neurons growing peripherally,
another by peripheral neurons growing centrally. These results suggest that the formation of
common nerves and neuronal pathfinding in the leech involves multiple sets of growth cone
guidance strategies and morphogenetic mechanisms that belie its apparent simplicity.
c
1996 Wiley-Liss, Inc.
Indexing terms: Hirudo medicinalis, embryogenesis, pioneer neurons, tubulin, fasciculation
A central issue in understanding the development of
nervous systems is to determine how patterned neuronal
connections and common nerve pathways are established.
The embryonic environment, although relatively simple, is
constantly changing, and thus multiple mechanisms for
guidance and formation of nerves are likely to be required.
Although a cell’s intrinsic properties may strongly influence the orientation of its initial extension (Acklin and
Nicholls, 1990), temporally and spatially regulated distributions of both cellular and extracellular guidance factors,
which subsequently direct the navigation of the growth
cone, have been characterized in a number of different
systems (Letourneau et al., 1992; Palka et al., 1992;
Goodman and Shatz, 1993; Jellies and Johansen, 1995). In
the developing nervous system of the grasshopper embryo,
pioneer growth cones use information obtained from glia,
G 1996 WILEY-LISS, INC.
epithelial cells, basal laminae, and “guidepost” neurons in
their path to set up a scaffold of axon tracts that are
themselves then used by later extending axons as substrates for directed migration (Bentley and Keshishian, 1982; Goodman et al., 1982; Bentley and O’Connor,
1992). Mammalian pioneer neurons have also been characterized (McConnell et al., 1989; Stainier and Gilbert, 1990;
Easter et al., 1993; Meissirel and Chalupa, 1994), suggesting that similar mechanisms may be employed among
Accepted December 15, 1995.
Diane M. Kopp’s present address: Department of Zoology, University of
Texas, Austin, TX 78712.
Address reprint requests to John Jellies, Department of Biological Sciences, Western Michigan University, Kalamazoo, Michigan 49008. E-mail:
john.jellies@ztwmich.edu
2
.J. JELLIES ET AL.
divergent species to generate the initial patterns of neuronal growth during early development.
We have addressed the issues of nerve formation and
axon guidance in a leech, Hirudo medicinalis, which has
particularly accessible embryonic stages and identified neurons and which has been used for many studies of neuronal
development and regeneration (Fernandez and Stent, 1982;
Muller et al., 1992; Jellies and Johansen, 1995). The central
nervous system (CNS) is comprised largely of a chain of
segmentally iterated ganglia (Muller et al., 19811, each of
which contains about 400 neuronal somata (Macagno,
1980). These ganglia are connected to each other via large
intersegmental tracts (connectives) and communicate with
the peripheral tissues via stereotyped nerve pathways.
Leeches also possess constellations of peripheral sensory
neurons that extend axons into the CNS (Derosa and
Friesen, 1981;Johansen et al., 1992; Gascoigne and McVean,
1993).Thus, the stereotyped nerve pathways contain mixed
populations of efferent and afferent projections, the extensions and interactions of which can readily be analyzed by
using CNS- and peripheral neuron-specific antibodies and
by intracellular dye filling of single identified cells. The
present studies analyze the normal sequence of events that
are involved in the formation of the major peripheral nerves
during the earliest stages of leech embryogenesis. Our
results suggest that neuronal pathfinding in the leech
involves multiple sets of growth cone guidance strategies
and morphogenetic mechanisms providing a more detailed
framework in which to interpret the results of studies
examining the cellular and molecular basis of pathfinding.
MATERIALS AND METHODS
Animals
Hirudo medicinalis leeches were obtained from a laboratory breeding colony. Breeding, maintenance, and staging
at 22-25°C were as described elsewhere (Fernandez and
Stent, 1982; Jellies et al., 19871, except that embryos were
maintained in water that was made as sterile-filtered
solutions of 0.0005% commercial sea salt (Instant Ocean),
wtlwt. Embryonic day 10 (El01 was characterized by the
first sign of a tail sucker, and E30 was the termination of
embryogenesis.
Dye filling
Embryos were dissected in leech saline that contained 8%
ethanol and were pinned epidermis upward in a saline-filled
well of a Sylgard iDow Corningbcoated slide. Cells in the
germinal plate were visualized by using transmitted light
and DIC optics on a fixed-stage Leitz microscope equipped
with epifluorescence. Central and peripheral neurons were
intracellularly filled with lucifer yellow (LY), as described
elsewhere (Jellies et al., 1987, 1992; Jellies and Kristan,
198813, 1991).
Immunocytochemistry
The results of this paper are based on the immunocytochemical labeling of more than 200 individual embryos,
each of which displays multiple segments in different stages
of development. Dissected embryos with or without dyefilled cells were fixed overnight in 4% paraformaldehyde in
0.1 M phosphate buffer (pH 7.4) in the cold. Preparations
were then rinsed extensively with phosphate buffered saline (PBS) before further processing. A monoclonal anti-
body (mAb)directed against acetylated tubulin (ACT; Sigma 1
was used to label central neurons and their axonal projections. The primary mAb was used at a dilution of 1:1,000 in
10% goat serum, 1% Triton X-100, and 0.001% sodium
azide in PBS. Incubations were carried out overnight on ice
with constant agitation. Lan3-2 (Zipser and McKay, 1981;
McKay et al., 1983) was used to label sensillar neurons in
fixed germinal plates, as described elsewhere (Johansen et
al., 1992). Secondary antisera used included horseradish
peroxidase (HRP)-conjugated goat anti-mouse (BioRad) at
a dilution of 1500-1: 1,000, Texas Red-conjugated rabbit,
anti-mouse IgGl secondary antibody (Cappel) at a dilution
of 1:250, and FITC-conjugated rabbit anti-mouse IgGzIl
(Cappel) at a dilution of 1 : l O O . Double labels using Lan3-2
Fig. 1. Stereomicrograph of axon tracts labeled by the antibodj
acetylated tubulin (ACT) forming the segmentally iterated peripheral
nerves. This micrograph is of a single, flattened hemisegment of an E l 4
embryo showing the projection patterns of nerves from the central
nervous system (CNS). The ACT antibody labels an epitope expressed
by all centrally located neurons, a subset of peripheral neurons, and a
repeated pattern of longitudinal muscles (double-headed arrows). The
four major nerve roots, AA (anterior-anterior),MA (medial-anterior 1,
DP (dorsal-posterior), and PP (posterior-posterior), arise as bifurcations from two nerve roots (A, anterior; P, posterior) and project many
finer nerves and branches in the developinggerminal plate. Three ofthe
major nerves innervate the ventral, lateral, and dorsal body walls,
whereas one of them (DPj rises out of the plane of focus and extensively
innervates the dorsal body wall. The DP nerve projects along a muscle
that inserts near the ventral and dorsal midline and later projects
through the lumen of the animal. During embryonic development,
axons in the AA and P P nerves project a short distance beyond the
edges of the germinal plate (white dotted line) onto the larval muscle.
Anterior is to the left. Scale bar = 100 wm.
Fig. 2. The ACT antibody labels the earliest axonal projections
within and from the CNS of an E9 embryo. Anterior is to the left in
these four panels. A The first cells to express the ACT epitope are
midline bipolar cells that define the position of the ventral midline
(large white arrow). Following the pioneering of rudimentary interganglionic and intraganglionic commissures, the first peripheral projection
is in the position of the future posterior nerve root (small black arrows).
Cell bodies become antibody positive only later in development. The
posterior projection at the far right has not yet penetrated the
ganglionic capsule, whereas the one immediately to its left has just
exited the ganglion. The CNS-derived projection in the position of the
anterior nerve root (small white arrows) consistently lagged behind
that of the posterior by about 2.5-5 hours. B: At slightly older stages, it
can be seen that the initial axonal projections from the CNS do not
prefigure the anterior and posterior nerve roots, hut rather, they form
each of the four major nerve pathways independently (arrowheads). As
predicted from earlier work on this system, the longest and first
projection corresponds to the DP nerve. C,D: The dorsal P cell (P,),
which pioneers the DP nerve, was injected with lucifer yellow (LY)and
double labeled with ACT antibody. C shows the LY-injected growth
cone (green)of the PI)cell; D is a double exposure of the preparation at
the same focal plane through different filter sets. The double label with
the ACT antibody confirms that the growth cone of the PD axon
corresponds to the most distal extent of the pioneering projection
(arrowhead). The P cell was dye filled at a stage just before it had
reached the edge of the germinal plate (asterisks). In the double
exposure, there are two ACT-positive axons. The farthest is the one also
filled with LY and represents the primary axonal projection; the shorter
of the two is from a different cell not injected with LY (white arrows).
Examining consecutive adjacent segments revealed that the second,
shorter, axon is projected from a P cell located in the adjacent anterior
segment. The axon from the P cell in the neighboring segment appears
to fasciculate with the pioneer axon of the injected P cell. This also
demonstrates that the double label seen here is not the result of any
overlap in spectral emissions (“breakthrough”) through the two different fluorescent filters. Scale bars = 50 ym.
Figure I
Figure 2
4
J. JELLIES ET AL.
and ACT were possible because Lan3-2 belongs to the IgGl
subtype and the ACT antibody to the 1gGz~
subtype. Double
labels were generated by sequential incubation in primary
and secondary antibodies. HRP-conjugated antibody complexes were visualized by reaction in 3,3'-diaminobenzidine
(BioRad, 0.03%) and HzOz(0.01%)for about 10 minutes.
Microscopy and image processing
Dye-filled cells were photographed with Ektachrome 100
or 400 HC or Ektar 25 Daylight film. High-definition stereo
light microscopy was performed by using the Edge microscope as described by Greenberg and Boyde (1993). Digital
images were obtained by using a high-resolution Paultek
cooled CCD camera, a Pixel Buffer framegrabber (Perceptics), and the NIH-Image software. The digital images were
pseudocolored and imported into Photoshop (Adobe), where
they were image enhanced and merged before being downloaded to a slide printer.
RESULTS
To examine the early development of axonal pathways,
we used a mAb directed against ACT. Tubulin antibodies
have been used to study the axonal projections of vertebrate
neurons (Black et al., 1989; Yaginuma et al., 1990; Wilson
and Easter, 1991) and labels all known central neurons and
their processes in the medicinal leech in addition to a
subpopulation of peripheral neurons. Figure 1 shows the
three-dimentional relationship of all the peripheral nerve
branches in an E l 4 embryo labeled with the ACT antibody.
In Hirudo, each midbody ganglion projects two major nerve
roots bilaterally that penetrate an enclosing sinus (Muller
et al., 1981; Sawyer, 1986). There is a single anterior (A)
and posterior (P) root in each hemisegment (Fig. 1).These
major roots bifurcate just as they enter the muscle of the
body wall, giving rise to four major nerves: AA (anterioranterior), MA (medial-anterior),DP (dorsal-posterior), and
P P (posterior-posterior; Ort et a]., 1974; Kretz et al., 1976;
Sawyer, 1986). Three of these (AA, MA, and PP) enter the
body wall and then branch extensively, whereas the fourth
nerve (DP) projects unbranched along a bundle of flattener
muscles that is within the body cavity and separate from
the body wall. This nerve projects directly toward the dorsal
body wall, where it then branches extensively in this region.
In addition to neurons, the mAb consistently labeled a
subset of longitudinal muscle fascicles (double-headed arrows, Fig. l),two in the ventralilateral region, two in the
dorsal region, and one in the lateral region of the body wall.
The ACT antibody recognizes axons and growth cones as
they are extended (see Fig. 2), whereas cell bodies only
become antibody positive at later stages of development.
Early CNS projections and nerve condensation
Studies of neuronal pathfinding have generally concentrated on either individually dye-filled cells or subsets of
cells labeled by selective antibodies. One of the assumptions
of such work is that the steering decisions of individual
growth cones seen early in development can be deduced by
the final stereotyped pattern of nerves and branches.
Nevertheless, these events have not generally been examined against a background of the total complement of
axonal projections, which would allow for the direct comparison of the spatiotemporal interactions between the different neuronal subpopulations. By using the ACT antibody to
label CNS efferents, we found that some aspects of the
earliest patterns established by axonal projections in Hirudo are transiently quite different from that of the adult
pattern of nerves (Fig. 2).
Individual embryos exhibit a rostral-caudal gradient of
development with approximately 2.5 hours between each
adjacent segment (Jellies and Kristan, 1991). Thus, examining several adjacent segments in single embryos yields a
relative sequence of axonal extension. As predicted from
earlier work on neuronal development in Hirudo, the first
staining revealed by ACT is seen on the midline bipolar cells
(large white arrow, Fig. 2A; McGlade-McCulloh et al.,
1990). This is followed in time (inferred from the relative
rostrocaudal position) by labeling of a limited number of
axons that project contralaterally to form one of the
commissures and axons that project intersegmentally within
the CNS. The earliest projections from the CNS are not
seen until after intersegmental axonal projections are continuous, and they exit the ganglionic primordium in a
position roughly corresponding to that of the future posterior root (small black arrows, Fig. 2A). This projection is
followed by an axonal extension in the position of the more
anterior root, which was typically one to two segments
more anterior (small white arrows, Fig. 2A) and thus
developing 2.5-5 hours after the posterior projections.
As predicted from earlier work on pioneer neurons in the
leech, the first projection of CNS efferents (Fig. 2A) corresponds to the formation of the DP pathway, whereas the
initial anterior projection corresponds to the MA nerve (Fig.
2B). The DP nerve in Hzrudo (Jellies et al., 199413) and
Haementeria (Kuwada, 1985) is pioneered by the dorsal
pressure-sensitive mechanosensory neurons (PDcells). Separate labels using ACT antibody and intracellular LY-filled
PD-cellaxons (Fig. 2C,D) demonstrate that the mAb labels
the first axonal projections along their entire length from
the cell body to the distal growth cone. Based on this result,
it appears reasonable to assume that the ACT antibody
serves as a marker for all the axonal extentions from the
CNS.
Although the initial projections are reminiscent of the
final pattern of nerves, examination of slightly older ganglia
revealed a phase of growth wherein the four nerve branches
(AA, MA, DP, and PP) were separate, projecting individually rather than arising from a bifurcation of growth cones
(Fig. 2B). Consequently, the apparent branch point is not
formed by pathway bifurcation but rather by a secondary
condensation of the individual tracts, which coalesce pairwise (Fig. 3; see also Fig. 2B). Thus, any early projections
that fasciculate with the pioneers of any single tract would
be restricted to grow along that particular tract, whereas
later projectinggrowth cones might have the opportunity to
send projections into each or choose one nerve branch over
another.
Peripheral neurons labeled
by the ACT antibody
In addition to labeling central neurons, the ACT antibody
labels a subset of peripheral neurons. However, early
embryos labeled with the ACT mAb did not reveal any of
these neurons that become ACT positive only during later
stages of development (Fig. 4). The identity and projections
of some of these peripheral cells were investigated by
intracellular dye-filling and double labels. There were several different classes of cells that labeled with ACT antibody
during later development, two of which have been previ-
MULTIPLE MECHANISMS OF NERVE FORMATION
Fig. 3. Formation of the proximal nerve roots by a
secondary condensation of initially parallel tracts. The
figure shows ACT-labeled hemisegments from comparable posterior regions on progressive embryonic days
( A ) 10, (B) 11, and (C) 14. Anterior is to the left. There is
an orderly progression of coalescing axonal tracts that
results in apparent points of bifurcation (small black
arrows). The condensation of nerves to form the two
major nerve roots begins at a relatively early stage, before
or just as the PD cell axon pioneering the DP nerve
(arrowheads show similar locations on the axon in A-C)
has reached the edge of the germinal plate (dotted arrows
in A-C). Although the condensation begins just outside
the ganglionic capsule, it moves peripherally and eventually comes to delineate the point at which nerves enter
the ventral body wall. At E l 4 (C), the ACT antibody
labeling begins to reveal several peripheral neurons. The
approximate locations of the major peripheral neurons
are indicated by small white arrows (a-d). These designations correspond to the locations of the cells shown in
Figure 4A-D. Scale bar = 50 pm.
Fig. 4. ACT-positive peripheral neurons are uniquely identifiable
cells. Anterior is to the left, and the relative positions of each neuron can
be seen in Figure 3. Each panel shows LY-injected neurons double labeled
with the ACT antibody. A A previously described stretch receptor in the
ventral body wall. This neuron projects a single large-caliber axon into
the CNS, where it forms a local terminal arbor. For reference, the cell
body of this neuron is indicated by the arrow here and in B and C. The
white brackets delineate the extent of flattened membranous veils
associated with longitudinal muscles, which are also labeled by the ACT
antibody. B: A single monopolar neuron reliably found just lateral to the
ventral stretch receptor. C: The previously identified nephridial nerve
cell innervates the developing nephridium, a portion of which can be seen
by nonspecific background staining highlighted by white dots. D One of
several peripheral bipolar cells found in stereotyped positions associated
with the major nerves. Scale bars = 25 pm.
5
MULTIPLE MECHANISMS OF NERVE FORMATION
ously identified (Wenning, 1983; Blackshaw, 1993) and two
hitherto unidentified neurons.
One of the peripheral cell types labeled was the putative
stretch receptors of the body wall, which have a characteristic association with specialized longitudinal muscles, project
large-diameter axons, and have elaborate characteristic
terminal branches within the CNS (Johansen et al., 1984;
Blackshaw, 1993). The most ventral of these (Fig. 4A)
arises along the AA nerve, with its cell body positioned
between two specialized longitudinal muscle fascicles, and
it has bipolar projections that expand as they cross the
muscle. These neurons begin to stain strongly with the
ACT antibody during later embryonic stages (E14-El6;
hollow white arrow, Fig. 4A-C). There is also a previously
unidentified, ACT-positive monopolar cell just lateral to the
stretch receptor along the AA nerve (Fig. 4B). Another
reliably stained peripheral neuron was the previously identified nephridial nerve cell (NNC; Wenning, 1983; Wenning
et al., 1993; Fig. 4C). These neurons are bilaterally paired
and project a large dendritic arbor that ramifies along the
nephridium and have a single axon extending toward the
CNS along the MA nerve. A fourth class of peripheral cell
was also routinely found (Fig. 4D) by using the ACT mAb.
The spindle-shaped bipolar cells are associated with the
major peripheral nerves and project long processes laterally
toward the edge of the germinal plate and medially toward
7
the ventral midline. Indeed, many of these processes extended beyond the margin of the germinal plate onto larval
muscle (e.g., note ACT-positive projections shown in Figs.
1, 3). These cells are likely to be neurons because they are
similar in appearance and location to cells labeled with the
mAb Lan3-8, which specifically labels all peripheral and
central neurons in hirudinid leeches (Johansen and Johansen, 1995). However, these bipolar cells' ventral projections
do not enter the ganglia. None of these four peripheral cell
types showed ACT labeling until well after the major nerve
tracts were established, and their possible involvement in
the formation of the earliest peripheral pathways could
therefore not be determined in this study.
Formation of common nerve pathways
by central and peripheral neurons
A major issue concerning peripheral nerve formation
relates to the relative temporal contributions from the CNS
and peripheral nervous system (PNS) in establishing common nerve pathways. Are they pioneered by either the CNS
or peripheral neurons or do both groups of neurons play a
role? The ACT antibody labels central neurons and some
peripheral neurons; however, a substantial population of
peripheral neurons arises early in embryogenesis that is not
labeled by the ACT antibody. These neurons, however, do
label with a different mAb, Lan3-2, which is specific for
sensillar and extrasensillar sensory neurons (McKay et al.,
1983; Johansen et al., 1992) that project axons toward the
Fig. 5. The S3 sensillum contains Lan3-2-positive afferents that
CNS early enough to make them possible candidates for
pioneer the MA nerve. Embryos at E8-E9 were double labeled by using
pioneering a t least one of the major nerves (Jellies et al.,
both ACT and Lan3-2 antibodies followed by an FITC-conjugated
1994b; Johansen et al., 1994; Jellies and Johansen, 1995).
secondary antibody (green) to localize the ACT epitope and a Texas
The
first population of the Lan3-2-positive peripheral neuRed-conjugated secondary antibody (red) to localize the Lan3-2 epitrons to differentiate is the sensillar neurons. These neurons
ope. The ventral midline is aligned with the far left-hand margin of each
are mixed sensory afferents that arise in seven bilateral
panel; anterior is up. A: Double exposure shows both labels in a
posterior segment. Solid white arrow indicates afferent axons (red) clusters aligned with the central annulus of each segment.
along the MA path, and hollow white arrow indicates efferent (green- They are designated Sl-S7, with S1 being the most ventral
yellow) axonal projections pioneering the DP nerve. B: Enhanced
and S7 the most dorsal. S1-S5 extend axons toward the
digitized image from a segment comparable to that shown in A in a
CNS
along the MA nerve, and S6 and S7 project along the
sibling embryo. The FITC and Texas Red labels were pseudocolored
DP nerve (Johansen et al., 1992). Previous studies using
green and red, respectively, and the small area of overlap between the
mAb Lan3-2 showed that the first sensillar neurons arise in
two labels is bright yellow. C: Double exposure shows both labels in a
more anterior segment of the same embryo shown in A. Within a few
S3 and extend growth cones directly toward the CNS,
hours followingthe peripheral ingrowth of the Lan3-2-positive sensillar where they segregate into distinct fascicles (Johansen et al.,
neurons in the MA path, there is a parallel axonal tract growing
1992). We examined the relative contribution of these
outward from the CNS. D: Enhanced digitized image from a segment
peripheral axons to the establishment of the nerves by
comparable to that shown in C in a sibling embryo shows both inward
simultaneously using the ACT mAb to label outgrowing
(solid arrow) and outward (hollow arrows) axonal projections as well as
the initial stages of S3 axon pathway selection within the CNS. In this
projections from the CNS and the Lan3-2 mAb to label the
figure, anterior is up and dorsal to the right. Scale bar = 25 pm.
ingrowing peripheral axons (Fig. 5).
Our ACT-labeled embryos (Fig. 2) show that, although
Fig. 6. Axons of Lan3-2-positive extrasensillar neurons migrate
the early posterior projections are present in young ganglialong select efferent pathways. Anterior is to the left. A: E l 4 embryo onic primordia, an ACT-positive anterior projection has not
labeled with Lan3-2 antibody shows the fascicles of sensillar axons that
yet' been extended (Fig. 5A,B). In these segments, simultaconstitute a portion of the MA nerve (contributed by SlLS5 sensillar
neurons) and DP nerve (contributed by S6 and S7 sensillar neurons). neously staining the sensillar neurons with Lan3-2 clearly
revealed that the growth cones of S3 neurons extended to
Only a few scattered extrasensillar neurons have differentiated at this
the ganglion, penetrated the outer layer of cell bodies, and
stage, as indicated by small black arrows. B: A sibling embryo of the
preparation shown in A is labeled with the ACT antibody to reveal all of
came into contact with the interior neuropil before any
the available efferent pathways at this stage. The DP nerve is out of the ACT-positive CNS efferents were directed peripherally
focal plane in this panel, and the asterisk indicates the position of the
(Fig. 5A,B). Thus, this anterior nerve pathway correspondnephridial bladderipore complex in A and B. C: Enhanced digitized
ing to the future MA nerve appears to be pioneered by S3
image of a peripheral region between the MA and AA nerves from an
E l 6 embryo prepared as a double label as in Figure 5. The asterisks axons rather than by CNS efferents as is the case for the DP
indicate Lan3-2-positive extrasensillar neuron somata. The pathways
nerve. A short time after reaching the central neuropil, S3
taken by the extrasensillar axons are highlighted by white arrows. Even
growth cones segregate along intersegmental tracts within
those neurons that arise immediately adjacent to the AA nerve migrate
the CNS (Johansen et al., 1992). At this stage, ACT-positive
along CNS efferents (green) to reach and fasciculate with other
projections from the CNS could be found coursing peripherLan3-2-positive axons in the MA nerve. Yellow indicates regions of
overlap between afferent and efferent axon pathways. Scale bars = 50 ally along the previously established Lan3-2-positive tracts
pm in A,B, 25 pm in C.
(Fig. 5C,D). At the earliest stages where central projections
8
J. JELLIES ET AL.
extending to the periphery were coincident with the peripheral projections centrally, they appeared to be aligned but
not intermixed with both labels clearly distinguished as
separate at the light microscopic level (Fig. 5C). Thus, the
fusion of the ACT-positive and Lan3-2-positive fascicles
into a common nerve pathway may also be the result of a
secondary morphogenetic process after independent pioneering events.
An interesting aspect of leech development is the finding
that numerous extrasensillar sensory neurons recognized
by the Lan3-2 antibody start to differentiate relatively late
in development at E l 6 and then continue to increase in
number throughout the life span of the leech (Peinado et
al., 1990; Johansen et al., 1992). From double labels with
Lan3-2 and ACT antibodies, it appears that these latedifferentiating neurons use the peripheral projections of
central neurons as a guide to reach the major nerve trunks,
where they then selectively fasciculate with the sensillar
axons (Fig. 6C). However, among the two anterior nerve
branches, only the MA nerve branch is used for such
tracking. Extrasensillar sensory neurons appearing right
on the AA nerve branch ignore these axons and project
instead only along axons coming from the MA nerve (Fig.
6C). This suggests that only a subset of central neurons
may have the capacity to serve as guides for the extrasensillar neurons and/or that some branches express repulsive
cues that result in these neurons’ axons being routed to
enter the CNS only through certain nerve pathways. Labeling of sibling embryos separately with Lan3-2 and ACT
antibodies shows that, at E l 6 when the extrasensillar
neurons differentiate, all the major peripheral projections
from the CNS are in place (Fig. 6A,B). Thus, the genesis of
extrasensillar neurons is developmentally synchronized
with the establishment of efferent peripheral projections,
which ensures that the proper guidance cues are in place
when they differentiate. Although not shown here, extrasensillar axons from the more posterior regions collect within
the PP nerve, forming a separate fascicle that enters the
ganglion through the posterior nerve root. These observations strongly suggest that the extrasensillar axons are
selecting particular axon tracts over others as they navigate
toward the CNS and that they use CNS efferents as guides
to reach the major nerve trunks.
DISCUSSION
There are potentially many different mechanisms by
which stereotyped nerve pathways can be established (Goodman and Shatz, 1993). In the medicinal leech, peripheral
neurons project axons toward the CNS and then segregate
into discrete tracts or form characteristic neuropilar arborizations. Likewise, axons extend first within, and later from,
the CNS into the periphery along stereotyped pathways. In
the present study, we have followed the initial formation of
the four major peripheral axonal pathways in Hirudo and
demonstrated that they were established in a spatially
discrete fashion. In addition, we have shown that the
apparent branch point seen in the mature leech, where the
proximal nerve roots each “bifurcate” to give rise to two
nerves, was actually formed by a secondary morphogenetic
mechanism involving a condensation of previously extended, parallel axon tracts. Thus, because the nerve tracts
were found to be formed individually, early navigation
along each should be considered independently. Furthermore, by having demonstrated that the ACT mAb stained
the earliest projecting CNS axons including their growing
tips, we have confirmed that the DP nerve is pioneered by
one of the P cells as previously shown by using other
techniques (Jellies et al., 1994b). In contrast, the MA nerve
was pioneered by sensory afferents from the S3 sensillum.
Consequently, in the leech, we found that neurons from
both the CNS and PNS contribute to the establishment of
common nerve pathways in the periphery. This situation is
similar to that described for the establishment of peripheral
nerves in the grasshopper limb (Ho and Goodman, 1982;
Keshishian and Bentley, 1983a,b) and Drosophila body
(Hartenstein, 1988). Although most of the emphasis has
been placed on the pioneer neurons arising peripherally, in
both systems there is also a population of central neurons
with peripherally directed growth cones that navigate at
least portions of their routes independently.
Although the mechanisms that give rise to the compaction of initially separate axonal pathways are not known,
another prominent example of neural condensation can be
seen in the formation of the subesophageal ganglion during
embryogenesis. Ganglionic primordia are initially very close
together with short connectives, followed by a period of
elongation, and, for a very brief period during embryogenesis (during E8-E9), the elongated connectives joining the
first four ganglia are visible (Fernandez and Stent, 1982).
In addition, our results may at least partly explain an
apparent species difference in the spatial and temporal
details of nerve formation between Hirudo and the glossiphoniid leech Helobdella (Braun and Stent, 1989a,b).In the
glossiphoniid, the MA and AA nerves remain distinct and do
not seem to undergo a comparable secondary condensation
(Braun and Stent, 1989a). This is indeed reminiscent of the
earliest situation in the formation of the nerves of Hirudo
and may represent a more primitive condition.
Our results here and in other studies (Jellies et al.,
1994a,b;Jellies and Johansen, 1995) have supported a role
for the interdependence of outgrowth from the CNS and
ingrowth in establishing the stereotyped nerves. As was the
case for the posterior nerves, the anterior nerve root was
not necessarily established by the S3 axons alone, and the
AA nerve might well be pioneered by as-yet-unidentified
neurons. The present study did not address the issue of
whether the S3 pioneers are necessary substrates for axons
extending from the CNS. However, double-labeled preparations from somewhat later developmental stages clearly
showed that later developing extrasensillar neurons preferentially extended along the Lan3-2-positive tracts that
previously established the MA nerve and not along the
Lan3-2-negative AA nerve. In the case of pioneer neurons
in insects, some pioneers are necessary for subsequent
navigation, and some are not (Edwards et al., 1981; Keshishian and Bentley, 1 9 8 3 ~ Bentley
;
and O’Connor, 1992). It
has been proposed that the “pioneering” of a pathway may
often be the result of other morphogenetic events and that
particular pioneers may not, therefore, be necessary for
subsequent navigation (Keshishian and Bentley, 1 9 8 3 ~ ) .
We might expect such mixed results in the leech as well.
Some pioneers may simply be the first neurons to navigate a
pathway that can be used by later neurons, whereas others
may be necessary to establish “labeled pathways” when the
environment is simple (Goodman et al., 1982).
We had previously used direct intracellular dye injection
to establish one of the P cells as the DP pioneer in Hirudo
(Jellies et al., 1994b).However, there are two pairs of P cells
in Hirudo (Muller et al., 1981). The dorsal P cell (PD)
MULTIPLE MECHANISMS OF NERVE FORMATION
projects ipsilaterally within the DP nerve, whereas the
ventral P cell (Pv) projects two peripheral axons, one in the
anterior nerve root. Although we had previously shown
that PD established the DP pathway, intracellular dye
filling of Pv (which does send one axon in the anterior
nerves) was more ambiguous, and its anterior peripherally
directed axon was always coextensive with the ingrowing
Lan3-2-positive S3 axons at the stages examined (Jellies et
al., 1994b). Our present study confirmed that, although
outgrowth in the MA nerve can occur early, it is preceded by
ingrowth of S3 peripheral growth cones. It is possible that
the ingrowing and outgrowing axons were utilizing the
same or closely apposed substrates rather than each other.
Such substrates might include cells that underlie incipient
pathways such as have been previously described in this
system (Kuwada, 1985; Jellies and Kristan, 1988b). Along
these lines, in another study (Jellies et al., 1994a) it was
shown that CNS-associated cues are required for the
continued directed navigation of Lan3-2-positive axons
toward the ventral midline. Thus, although one major
peripheral nerve is pioneered by sensillar afferents and
another by CNS outgrowth, two others (AAand PP) have
not been examined in this light, and there remain as-yetunidentified cues associated with the CNS that are also
involved in the directed navigation of neuronal growth
cones. This complexity in what at first may appear to be a
simple system reveals that the formation of common nerves
and neuronal pathfinding in the leech involves multiple sets
of growth cone guidance strategies and morphogenetic
mechanisms that may be of general significance.
ACKNOWLEDGMENTS
We thank Evanne Maher and Edge Scientific Instrument
Corporation (Los Angeles, CA) for graciously providing us
access to their newly developed real-time three-dimensional
microscope and for generating the stereo pair used in
Figure 1. Work reported here was supported by NSF grants
9609701 ( J . Jellies), DIR 9113595 (J.Johansen), and NIH
grant 28857 ( J . Johansen). J. Jellies is a Fellow of the
Alfred P. Sloan Foundation. Journal paper 5-16745 of the
Iowa Agriculture and Home Economics Experiment Station, Ames, Iowa, Project 3371, was supported by Hatch Act
and State of Iowa funds.
LITERATURE CITED
Acklin, S.E., and J.G. Nicholls i1990)Intrinsic and extrinsic factors influencingproperties and growth patterns of identified leech neurons in culture.
J. Neurosci. 10:1082-1090.
Bentley. D., and H. Keshishian (19821 Pathfinding by peripheral pioneer
neurons in grasshoppers. Science 218:1082-1088.
Bentley, D., and T.P. O’Connor (1992) Guidance and steering of peripheral
pioneer growth cones in grasshopper embryos. In P.C. Letourneau, S.B.
Kater. and E.R. Macagno (eds): T h e Nerve Growth Cone. New York:
Raven Press, pp. 265-282.
Black, M.M., P.W. Baas, and S. Humphries (1989) Dynamics of wtubulin
deacetylation in intact neurons. J. Neurosci. 9:358-368.
Blackshaw, S.E. (1993) Stretch receptors and body wall muscle in leeches.
Comp. Biochem. Physiol. 105A:643-652.
Braun, J., and G.S. Stent (1989a) Axon outgrowth along segmental nerves in
the leech. I. Identification ofcandidate guidance cells. Dev. Biol. 132.471485.
Braun, J . , and G.S. Stent (1989b)Axon outgrowth along segmental nerves in
the leech. 11. Identification of actual guidance cells. Dev. Biol. 132.486501.
Derosa, Y.S., and W.O. Friesen (1981)Morphologyofleech sensilla: Observations with the scanning electron microscope. Biol. Bull. 160:383-393.
9
Easter, S.S. Jr., L.S. Ross, and A. Frankfurter (1993)Initial tract formation
in the mouse brain. J. Neurosci. 13285-299.
Edwards, J.S., S.-W. Chen, and M.W. Berns (1981) Cercal sensory development following laser microlesions of embryonic apical cells in Acheta
domesticus. Dev. Biol. 1321448457.
Fernandez, J . , and G.S. Stent (1982) Embryonic development of the
hirudinid leech Hirudo medicinalis: Structure, development and segmentation of the germinal plate. J. Embryo]. Exp. Morphol. 72.71-96.
Gascoigne, L., and A. MrVean (1993) Postembryonic growth of two peripheral sensory systems in the medicinal leech, Hirudo medicinalis. Biol.
Bull. 185.388-392.
Goodman, C.S., and C.J. Shatz (1993) Developmental mechanisms t h a t
generate precise patterns of neural connectivity. Cell 7277-98.
Goodman, C.S., J.A. Raper, R.K. Ho, and S. Chang (1982) Pathfinding by
neuronal growth cones in grasshopper embryos. In S. Subtelny and P.
Green (eds):Developmental Order: Its Origin and Regulation. New York:
Alan R. Liss, pp. 275-3165.
Greenberg, G., and A. Boyde i1993) Novel method for stereo imaging in light
microscopy a t high magnifications. Neuroimage 1.121-128.
Hartenstein, V. i1988) Development of Drosophila larval sensory organs:
Spatiotemporal pattern of sensory neurones, peripheral axonal pathways
and sensilla differentiation. Development 102:869-886.
Ha, R.K., and C.S. Goodman i1982)Peripheral pathways are pioneered by a n
array of central and peripheral neurones in grasshopper embryos.
Nature 297.404-406.
Jellies, J . , and J . Johansen (19951 Multiple strategies for directed growth
cone extension and navigation of peripheral neurons. J . Neurobiol.
27.310-325.
Jellies, J., and W.B. Kristan, J r . (1988a)Embryonic assembly of a complex
muscle is directed by a single identified cell in the medicinal leech. J.
Neurosci. 8.3317-3326.
Jellies, J . , and W.B. Kristan, Jr. (1988a) An identified cell is required for the
formation of a major nerve during embryogenesis in the leech. J.
Neurobiol. 19: 153-165.
Jellies, J . , and W.B. Kristan, J r . (1991) The oblique muscle organizer in
Hirudo medicinalis, a n identified embryonic cell projecting multiple
parallel growth cones in a n orderly array. Dev. Biol. 148.334-354.
Jellies, J., C.M. Loer, and W.B. Kristan, J r . (1987) Morphological changes in
leech Retzius neurons after target contact during embryogenesis. J.
Neurosci. 7:261%2629.
Jellies, J.,D.M. Kopp, and J.W. Bledsoe (1992) Development of segment- and
target-related neuronal identity in the medicinal leech. J. Exp. Biol.
170.71-92
Jellies, J., K. Johansen, and J. Johansen i1994a) CNS-associated cues a r e
required for centrally-directed navigation by peripheral sensory neurons
in embryonic leech. Sac. Neurosci. Abstr. 20.1064.
Jellies, J., K. Johansen, and J . Johansen (1994b) Specific pathway selection
by the early projections of individual peripheral sensory neurons in the
embryonic medicinal leech. J . Neurobiol. 25.1187-1199.
Johansen, J., S. Hockfield, and R.D.G. McKay (1984) Axonal projections of
mechanosensory neurons in t h e connectives and peripheral nerves of the
leech, Haemopis marmorata. J . Comp. Neurol. 226:255-262.
Johansen, J . , K.M. Johansen, K.K. Briggs, D. Kopp, and J . Jellies (19941
Hierarchical guidance cues and selective axon pathway formation of
sensory neurons. In F.J. Seil (ed): Progress in Brain Research. Amsterdam: Elsvier, pp. 109-120.
Johansen, K.M., and J . Johansen (1995) Filarin, a novel invertebrate
intermediate filament protein present in the axons and perikarya of
developing and mature leech neurons. J. Neurobiol. 27227-239.
Johansen, K.M,, D.M. Kopp, J. Jellies, and J. Johansen (1992) Tract
formation and axon fasciculation of molecularly distinct peripheral
neuron subpopulations during leech embryogenesis. Neuron 8:559-572.
Keshishian, H., and D. Bentley (1983a) Embryogenesis of peripheral nerve
pathways in grasshopper legs I. The initial nerve pathway. Dev. Biol.
96: 89- 102.
Keshishian, H., and D. Bentley (198313)Embryogenesis of peripheral nerve
pathways in grasshopper legs 11. The major nerve routes. Dev. Biol.
96:103-115.
)
of peripheral nerve
Keshishian, H., and D. Bentley ( 1 9 8 3 ~Embryogenesis
pathways in grasshopper legs 111. Development without pioneer neurons.
Dev. Biol. 96:116-124.
Kretz, J.R., G.S. Stent, and W.B. Kristan, J r . (1976) Photosensory input
pathways in the medicinal leech. J . Comp. Physiol. 106:l-37.
Kuwada, J.Y. (19851 Pioneering and pathfinding by an identified neuron in
the embryonic leech. J. Embryo]. Exp. Morphol. 86.155-167.
10
Letourneau, P.C., S.B. Kater, and E.R. Macagno (1992) (eds): The Nerve
Growth Cone. New York: Raven Press.
Macagno, E.R. (1980) Number and distribution of neurons in leech segmental ganglia. J Comp. Neurol. 190.283-302.
McConnell, S.K., A. Ghosh, and C.J. Shatz (1989) Subplate neurons pioneer
the first axon pathway from the cerebral cortex. Science 245:978-981.
McGlade-McCulloh, E., K.J. Muller, and B. Zipser (1990) Expression of
surface glycoproteins early in leech neural development. J. Comp.
Neurol. 299.123-131.
McKay, R.D.G., S. Hockfield, J. Johansen, I. Thompson, and K. Frederiksen
(1983) Surface molecules identify groups of growing axons. Science
222788-794.
Meissirel, C., and L.M. Chalupa (1994) Organization ofpioneer retinal axons
within the optic tract of the rhesus monkey. Proc. Natl. Acad. Sci. USA
91:3906-3910.
Muller, K.J., J.G. Nicholls, and G.S. Stent (1981) (eds): Neurobiology of the
Leech. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory.
Muller, K.J., X. Gu, E. McGlade-McCulloh, A. Mason, and S.R. Young (1992)
The growth cone: Sprouting and synapse regeneration in the leech. In
P.C. Letourneau, S.B. Kater, and E.R. Macagno (eds): The Nerve Growth
Cone. New York: Raven Press, pp. 453462.
Ort, C.A., W.B. Kristan, J r . , and G.S. Stent (1974) Neuronal control of
swimming in the medicinal leech 11. Identification and connections of
motor neurons. J. Comp. Physiol. 94:121-154.
J. JELLIES ET AL.
Palka, J., K.E. Whitlock, and M.A. Murray (1992) Guidepost cells. Curr
Opin. Neurobiol. 248-54
Peinado, A., B. Zipser, and E.R. Macagno (1990) Segregation of afferect
projections in the central nervous system of the leech Hirudo medicinclis. J. Comp. Neurol. 301332-242.
Sawyer, R.T. (1986) Leech Biology and Behavior. Oxford: Clarendon Press.
Stainier, D.Y.R., and W. Gilbert (1990) Pioneer neurons in the mouse
trigeminal sensory system. Proc. Natl. Acad. Sci. USA 87.923-927.
Wenning, A. (1983) A sensory neuron associated with the nephridia of the
leech Hirudo medicinalis L. J. Comp. Physiol. 152455-458.
Wenning, A., M.A. Cahill, U. Greisinger, and U. Kaltenhauser (1993)
Organogenesis in the leech: Development of nephridia, bladders and
their innervation. Roux Arch. Dev. Biol. 202329-340.
Wilson, S.W., and S.S.Easter, J r . (1991) Stereotyped pathway selection by
growth cones of early epiphysial neurons in the embyonic zebrafish.
Development 112.723-746.
Yaginuma, H., T. Shiga, S. Homma, R. Ishihara, and R.W. Oppenheirr
(1990) Identification of early developing axon projections from spinal
interneurons in the chick embryo with a neuron specific p-tubulir
antibody: Evidence for a new “pioneer” pathway in the spinal cord
Development 108.705-716.
Zipser, B., and R. McKay (1981) Monoclonal antibodies distinguish identifiable neurones in the leech. Nature 289:549-554.