Development 99, 173-186 (1987)
Printed in Great Britain © The Company of Biologists Limited 1987
173
Dorsal root ganglia grafts stimulate regeneration of denervated urodele
forelimbs: timing of graft implantation with respect to denervation
BRUCE L. TOMLINSON and ROY A. TASSAVA
The Ohio State University, Department of Zoology, 1735 Neil Avenue, Columbus, Ohio 43210-1293, USA
Summary
Amphibian forelimb regeneration is a nervedependent process; nerves presumably release one or
more neurotrophic factors that stimulate blastema cell
division. To date several candidate molecules/factors
have been shown to stimulate macromolecular synthesis and/or mitosis but sustained cell cycle activity
and blastema development have not been achieved.
Because dorsal root ganglia (DRG) implants are
capable of promoting regeneration of denervated
adult newt limbs (Kamrin & Singer, 1959), we have
evaluated the DRG stimulation of regeneration in
denervated limbs of adult newts and larval axolotls;
two alternative timing strategies were tested as a step
toward defining bioassay parameters that best reflect
neurotrophic activity. The frequency of regeneration
in denervated adult newt limbs was compared after
providing DRG before or at the time of denervation (to
maintain neurotrophic and cell cycle activity) versus
DRG implantation at various postdenervation times
(to resupply neurotrophic activity and restimulate
suppressed cell cycle activity).
The results show that denervated adult newt limbs
regenerated most frequently using the maintenance
strategy, but as the denervation interval was extended
in the restimulation strategy, the frequency of regeneration declined. Larval axolotl limbs responded positively in both maintenance and restimulation DRGgrafting protocols. These results suggest that the
efficacy of DRG stimulation of regeneration in adult
newts was related to the relative number of blastema
cells present at the time of denervation and the
proliferative status of the blastema cells; bioassays
with denervated adult newt limbs should be designed
with these constraints in mind. Because such constraints are not as problematic with the larval axolotl,
this species may provide the best opportunity for
further defining bioassay parameters related to the
neurotrophic stimulation of regeneration.
Introduction
(Globus, 1978; Maden, 1978) and/or G2 (Tassava &
Mescher, 1975) phases of the cell cycle. Appropriately therefore, DNA and protein synthesis as well as
mitotic and [3H]thymidine labelling indices have been
used as bioassay parameters to evaluate the efficacy
of treating denervated limbs with putative neurotrophic factors. In fact, increases in blastema macromolecular synthesis, or mitotic and DNA labelling
indices have been reported following treatment with
dorsal root ganglia (Globus & Vethamany-Globus,
1977; Vethamany-Globus, Globus & Tomlinson,
1978; Vethamany-Globus, Globus, Darch, Milton &
Tomlinson, 1984; Tomlinson, Globus & VethamanyGlobus, 1981), nerve extracts (Lebowitz & Singer,
1970; Singer, Maier & McNutt, 1976; Jabaily &
Singer, 1977; Carlone & Foret, 1979; Mescher & Loh,
Regeneration of the urodele forelimb is dependent
upon an adequate nerve supply (reviewed by Singer,
1952; Wallace, 1981). While the precise biochemical
nature of the neurotrophic influence and mechanism
of nerve action are as yet not defined, progress has
been reported towards definitive resolution of these
problems. Subsequent to denervation, macromolecular synthesis (Dresden, 1969; Lebowitz & Singer,
1970; Singer & Caston, 1972) and mitosis (Singer &
Craven, 1948; Mescher & Tassava, 1975; Globus,
1978; Tomlinson, Globus & Vethamany-Globus,
1984) are suppressed and it has been speculated that
the neurotrophic influence promotes sustained cell
cycling by stimulating events in either the Gl
Key words: newt, axolotl, regeneration, denervation,
dorsal root ganglia, graft implantation.
174
B. L. Tomlinson and R. A. Tassava
1981) and purified molecules (Mescher & Gospodarowicz, 1979; Globus, Vethamany-Globus, Kesik &
Milton, 1983; Mescher & Muniam, 1984; Carlone &
Rathbone, 1985).
The ultimate demonstration of neurotrophic activity will be the promotion of blastema development
on a denervated limb (Wallace, 1981; Olsen, Barger
& Tassava, 1984). Since blastema growth clearly
requires the coordinated syntheses of DNA, RNA
and proteins which culminate in mitotic divisions, it is
understandable that intermittent treatment of denervated limbs with purified molecules and/or extracts
has at present resulted in only partial recovery of
these parameters; no single chemical treatment has
resulted in growth of a regenerate (reviewed by
Wallace, 1981). In this regard it is noteworthy that
dorsal root ganglia (DRG), isolated from their central nervous system connections and grafted into
denervated limbs, are capable of stimulating blastema growth and regeneration (Kamrin & Singer,
1959). As a potentially complete source of neurotrophic factor(s) ganglia therefore provide the
opportunity to examine the relevance of the various
bioassay parameters currently employed. In the Kamrin and Singer study, DRG were excised between 10
and 13 days postamputation (thereby denervating the
limb) and simultaneously autografted to the distal
limb tissues. These early regenerates were technically
never denervated since the implanted DRG presumably maintained uninterrupted neurotrophic activity.
More recent studies using nerve extracts or isolated
molecules have not followed the timing protocol used
by Kamrin and Singer. Instead, treatment was
administered 48 h or more after denervation, at times
when macromolecular synthesis (Singer & Caston,
1972) and mitosis (Singer & Craven, 1948) were
already reduced. The reported small increases in
bioassay parameters (reviewed by Wallace, 1981)
therefore represent only minimal stimulation by
neurotrophic activity. Perhaps the introduction of an
interval between denervation and treatment with a
putative neurotrophic factor is responsible for the
minimal stimulations that have been observed.
In the present study we used the DRG-implantation technique (Kamrin & Singer, 1959) to investigate two alternative timing strategies with respect to
the neurotrophic stimulation of regeneration in
denervated limbs. We tested the hypothesis that as
the denervation interval increases, the regenerates
become committed to an alternative fate (i.e. differentiation or resorption) and no longer respond to
neurotrophic stimulation. We compared the frequency of regeneration in denervated limbs after
providing DRG implants before, or at the time of,
denervation (maintenance of neurotrophic and cell
cycle activity) versus implantation at various post-
denervation times (resupplying neurotrophic activity
to restimulate suppressed cell cycle activity). The
results showed that denervated adult newt limbs
regenerated most successfully using the maintenance
strategy and that as the length of the denervation
interval was extended to 1 and 2 days, successful
DRG stimulation of regeneration progressively declined.
Neither adult newt nor larval axolotl limbs regenerate after denervation; however, axolotl limbs are
unique in that following regrowth of the transected
nerves, limbs spontaneously regenerate without
reamputation (Schott6 & Butler, 1941; Petrosky,
Tassava & Olsen, 1980). Ingrowing nerve fibres
stimulate cells previously arrested in the cell cycle
(Olsen et al. 1984) but more detailed cell cycle studies
are impractical since the timing of reinnervation
exhibits considerable variability (10-14 days after
denervation; Olsen et al. 1984). DRG grafts, previously untested in denervated larval axolotl limbs,
provided a means of specifically timing neurotrophic
stimulation. We therefore examined the response of
denervated larval axolotl limbs to xenografts of adult
newt DRG, again using maintenance and restimulation strategies. DRG were grafted either concomitant with denervation (maintenance) or following
various denervation intervals (restimulation). Unlike
adult newt limbs, denervated larval axolotl limbs
responded positively to both maintenance and restimulation DRG-grafting protocols.
Materials and methods
Larval axolotls (Ambystoma mexicanum), of 50-60 mm
body length, and adult newts (Notophthalmus viridescens)
were anaesthetized in neutralized 0-15 % MS 222 and bilaterally amputated through either the proximal third of the
radius and ulna (larval axolotls) or distal humerus (adult
newts). Due to differences in the size of forelimbs and
hindlimbs, only forelimbs were used, thereby providing a
built-in control for size within each experimental series.
Since nerve regeneration is very rapid in the larval axolotl
(10-14 days, Olsen etal. 1984) compared to adult newts (14
days, Salley & Tassava, 1981), more distal amputation
planes were used in axolotls in order to increase the time
period over which the axolotl limbs remained denervated.
Limbs were denervated by transection of the third, fourth
and fifth brachial nerves and redenervated at 8-day intervals in axolotls and at 2-week intervals in newts. The
completeness of denervation was confirmed for every case
by the total absence of motor and sensory function, by the
failure of all denervated limb stumps to form a regeneration
blastema and by the presence of only degenerated nerve
fragments in histological sections of denervated limbs
stained for nerves according to the method of Samuel
(1953).
DRG stimulation of denervated urodele forelimb regeneration
Implant procedure
Dorsal root ganglia (the paired 3rd and 4th brachial and the
16th and 17th crural), pituitary glands, flank muscle or liver,
obtained from decapitated newts, were temporarily (15 min
maximum) stored in chilled Holtfreter's solution. A small
ventral incision was made in the limb skin at a level just
proximal to the original plane of amputation and, using a
single tyne of a pair of watchmaker's forceps, a tunnel was
created through the underlying tissues to the distal end of
the limb. Either control tissue or randomly selected
brachial or crural ganglia were then inserted into the tunnel
and positioned under the margin of the skin and wound
epithelium. The DRG were obtained solely from newts so
that a single source of neurotrophic factors could be
evaluated.
Adult newt control limbs received pituitary gland homografts and larval axolotl control limbs were implanted with
ganglion-sized newt flank muscle or liver. Additional control limbs were either denervated and sham implanted or
remained normally innervated. The different implantation
timing protocols used in each experiment are illustrated in
Fig. 1.
Differences in the timing of denervation and DRG
implantation were required in order to compensate for
different rates of limb and nerve regeneration between
these species. The time frame of each experimental series
was adjusted to compensate for the aforementioned faster
rate of nerve regeneration and the fact that innervated
limbs of larval axolotls can produce a blastema in 6-8 days
postamputation whereas adult newts require 14-18 days to
Maintenance
Restimulation
Series I
I0 1
Series II
0
Series III
8
10 11 12
14
12 14 15 16
Series IV
0h
SeriesV
I—
Denervation
Fig. 1. Timing protocol of experiments designed to test
the maintenance and restimulation strategies for the
promotion of regeneration on denervated urodele limbs
by grafted dorsal root ganglia (DRG). Adult newt
forelimbs were denervated (represented by the dotted
vertical line) 1 (Series I), 10 (Series II) or 14 days (Series
III) after amputation. Larval axolotl limbs were
denervated 1 (Series IV) or 4 (Series V) days after
amputation. In the maintenance strategy, DRG were
grafted either 2 days before or at the time of denervation
(represented by the number on, or to the left of, the
vertical line) and in the restimulation strategy, DRG were
grafted on various days after denervation (represented by
numbers to the right of the vertical line).
175
form a similar stage regenerate (Tomlinson, Goldhamer,
Barger & Tassava, 1985). If limbs are denervated at the
time of amputation (both in newts and larval axolotls)
limited cell division occurs for only a brief period of time
(Mescher & Tassava, 1975; Maden, 1978; Olsen et al. 1984).
Limbs denervated during the process of regeneration generally experience one of two fates (Singer & Craven, 1948;
Butler & Schotte, 1949); with relatively early denervations
the blastema resorbs and the limb fails to regenerate, but
when denervation is delayed to the midbud stage or later,
small hypomorphic spikes or small but complete limbs may
result (a phenomenon known as nerve-independence).
Denervation of blastema-stage regenerates results in a
gradual reduction of cell division over a 2-day interval
(Tassava & McCullough, 1978; Tomlinson et al. 1984).
Timing of DRG homografts into newt forelimbs
To examine the response of different-sized blastema cell
populations to the stimulation by homografted DRG, limbs
were implanted concomitant with denervation, either 1
(Series I), 10 (Series II) or 14 (Series IE) days after
amputation (Fig. 1).
To evaluate maintenance versus restimulation strategies,
a pair of DRG were implanted at different times with
respect to denervation in Series II and III (Fig. 1). In the
maintenance strategy DRG were implanted either 48 h
before or concomitant with denervation whereas in the
restimulation strategy, DRG were grafted 24, 48 and in
some cases 96 h after denervation.
Timing of DRG xenografts into axolotl forelimbs
Series TV
To maintain uninterrupted neurotrophic activity, a single
DRG was grafted at the time of denervation (24 h after
amputation). In denervated axolotl limbs dedifferentiated
cells participate in limited cell division 4 to 6 days after
amputation, whereas between 7 and 10 days postdenervation, proliferative activity is reduced to basal levels
(Maden, 1978; Olsen et al. 1984). DRG were implanted 4
days after denervation (day 5 postamputation) in an attempt to maintain the aforementioned limited cell cycle
activity. To restimulate cell division, DRG were implanted
8 days after denervation (9 days postamputation).
Series V
Axolotl limbs were denervated 4 days after amputation and
a pair of DRG were xenografted concomitant with (maintenance strategy), or 4 days after denervation (restimulation strategy).
Staging and histology
Progressive changes were monitored by staging the regenerates (Iten & Bryant, 1973) at 2-day intervals. To compare
with Kamrin & Singer (1959), limbs that regenerated to at
least the midcone stage within the time constraints of the
experiment (35 days postamputation for newts and 16 days
postamputation for axolotl limbs) were tabulated as successful regenerates. The frequency of regeneration between
implant groups was compared using a Chi-squared test for
the equality of binomial proportions.
176
B. L. Tomlinson and R. A. Tassava
Table 1. Regenerative responses observed in adult newt control limbs 35 days postamputation
Group
n
Nonregenerate
Spike
Series W
(Day 1 denervation)*
Innervated
Denervated
10
10
10
—
10
10
10
—
8
9
8
9
—
—
12
10
6
4
Midcone
Late
cone
Palette
3-4
digit
—
3
3
4
3
3
3
1
—
2
—
10
Series II
(Day 10 denervation)*
Innervated
Denervated
Pituitary graft
48 h before
48 h after
Series III
(Day 14 denervation)*
Innervated
Denervated
Limbs with midcone stage outgrowths, or better, were scored as regenerates,
* Time of denervation after amputatioii.
At the end of the experimental period, four or more
limbs per group were fixed in Bouin's fluid, embedded in
paraffin and longitudinally sectioned at 10/mi. One set of
representative sections from each limb was stained with
Delafield's haematoxylin and counterstained with eosin,
and to confirm denervation a second group of representative sections was stained for nerves (Mescher & Tassava,
1975; Olsen & Tassava, 1984) by the Samuel (1953) technique.
Results
Homografts to denervated newt limbs
Controls
10 days after amputation innervated limbs showed no
visible outgrowth, but mound-shaped regenerates
were present 14 days after amputation and by 35 days
regenerates ranged from midcone to digit stages
(Table 1); the range of stages observed is a consequence of normal variation (Iten & Bryant, 1973).
No regeneration was observed when control limbs
were denervated 1 or 10 days after amputation.
However, because one consequence of increasing
the postamputation age of the regenerates prior to
denervation is the accumulation of a larger blastema
cell population, denervation 14 days postamputation
permitted the subsequent development of several
very small spike regenerates (Table 1). In Series III
some regenerates had therefore become nerve-independent (Singer & Craven, 1948) at the time of
denervation (14 days postamputation). Comparable
tiny amorphic spikes, also observed on some of the
ganglia implanted limbs of Series III, were not
counted as DRG-stimulated regenerates.
Data in Table 1 and Fig. 2 also show that none of
the denervated limbs regenerated in either of the
pituitary-implanted control groups. The voracious
appetite, smooth moist skin and darkened pigmentation indicated that the ectopically located pituitary
glands continued to function (Tassava, 1969). Therefore, the stimulation of regeneration in denervated
newt limbs (reported below) can be directly attributed to the influence of grafted DRG.
Maintenance strategy
Implanting DRG either prior to, or at the time of,
denervation (1, 10 or 14 days postamputation) allowed the maintenance strategy to be examined in
adult newt limbs that contained sequentially larger
blastema cell populations at the time of nerve withdrawal. Limbs denervated and concomitantly implanted with DRG 1 day after amputation (Series I)
did not regenerate, whereas in Series II, DRG
grafted 48 h before or at the time of denervation (day
10), stimulated regeneration on 32 % arid 34 % of the
limbs, respectively (Table 2, Fig. 2). In Series III the
percentage of regenerating limbs increased further to
64 % and 57 % when DRG were homografted either
48 h before or at the time of denervation (day 14)
respectively (Fig. 3). Therefore, with each increase in
the interval between amputation and denervation a
significantly greater percentage of limbs (/ ) <0-05)
was stimulated to regenerate using the maintenance
strategy. Furthermore, in both Series II and III
implanting DRG 48 h prior to denervation was as
effective (P>005) as implanting DRG concomitant
with denervation.
DRG stimulation of denervated urodele forelimb regeneration
100
c
o
(10)
Restimulation strategy
80
(n)
"3 60
§
c 40
0
111
8
12
Pituitaries
12
10
11
14
Ganglia
Days postamputation
Fig. 2. A histogram showing the percentage of adult
newt limbs that regenerated in Series II (denervation 10
days postamputation; scored 35 days postamputation).
The open histogram bar illustrates that all of the
innervated (7) control limbs regenerated while the solid
bars show that none of the denervated (D) and pituitary
implanted control limbs regenerated. Cross-hatched bars
represent the percentage of DRG-implanted limbs
stimulated to regenerate. With increased intervals
between denervation (day 10) and subsequent
implantation (days 11, 12 and 14) the percentage of limbs
stimulated to regenerate progressively declined. The
number (n) of limbs in each group is given in
parentheses.
(n)
12
14
15
lf>
D
Days postamputation
Fig. 3. A histogram showing the percentage of adult
newt limbs that regenerated in Series III (denervated 14
days postamputation). While all of the innervated limbs
(/; open bar) regenerated, denervated controls (D; solid
bar) did not. A significantly greater percentage of limbs
(P<0-05) regenerated when ganglia were grafted (crosshatched bars) using the maintenance (days 12 and 14)
versus the restimulation (days 15 and 16) strategy. The
number (n) of limbs in each group is given in
parentheses.
To examine the effects of short denervation intervals
on the ability of adult newt limbs to respond to
neurotrophic restimulation, DRG were grafted 24, 48
or 96 h after nerve withdrawal. Data in Figs 2 and 3
illustrate a trend of declining stimulation by DRG
with increased periods of denervation. In Series II,
the frequency of stimulation was 25%, 15% and
12% with 24, 48 and 96 h intervals between denervation (day 10) and DRG grafting, respectively. In
Series III, when the ganglia were implanted 24 or 48 h
after denervation (day 14), only 18 % and 14 % of the
limbs regenerated, respectively. By increasing the
interval between amputation and denervation (i.e.
Series II versus III), there was no significant difference (P > 0-70) in the percentage of denervated limbs
stimulated by DRG to regenerate using the restimulation strategy. Similarly, within each series, the
percentage of limbs stimulated to regenerate declined
as the interval between denervation and subsequent
DRG implantation increased; however these declines
were not statistically significant (P> 0-40).
Comparison of the maintenance and restimulation
strategies
When examined in terms of the appropriate timing of
DRG implantation, the data in Figs 2 and 3 illustrate
that denervated adult newt regenerates respond more
frequently using the maintenance versus restimulation
strategy. There was no significant difference
(P = 0-10) in the response to maintenance and restimulation strategies in Series II. In Series III the use
of the maintenance strategy resulted in a significantly
greater percentage (P< 0-001) of regenerating limbs
compared to the results obtained using the restimulation protocol. The frequency of regenerate outgrowth was significantly greater compared to the
denervated controls of each series (either with or
without homografted pituitary glands) when DRG
were implanted 48h before (Series II, P = 0-04;
Series III, P = 0-001) or at the time of "denervation
(Series II, P = 0-03; Series III, P = 0-003). On the
other hand, it is important to note that compared to
denervated controls the frequency of regenerate outgrowth was not significantly stimulated when DRG
were grafted 24h (Series II, P = 0-08; Series III,
P = 0-16), 48h (Series II, P = 0-20; Series III,
P = 0-21), or 96h (P = 0-25) after denervation.
The morphological stage attained also depended
on the interval between denervation and DRG
implantation. More advanced regenerates were observed when there was no interruption of neurotrophic stimulation and conversely, less-advanced
regenerates were observed by prolonging the preimplantation denervation interval (Table 2). Series
178
B. L. Tomlinson and R. A. Tassava
Table 2. Stimulation of adult newt forelimb regeneration by DRG homografts
Implant group
n
Nonregenerate
Spike
Midcone
Late
cone
Palette
digit
10
10
—
—
—
—
—
28
29
19
18
—
1
2
—
—
5
—
3
7
2
24
20
8
18
17
6
—
—
1
1
1
1
2
1
—
1
—
—
2
1
—
14
14
5
5
—
1
3
4
3
1
1
1
2
2
11
14
6
10
3
2
—
—
—
2
2
—
—
—
3-4
Series IV
(Day 1 denervation)*
Maintenance
At time of
Series 11
(Day 10 denervation)*
Maintenance
48 h before
At time of
Restimulation
24 h after
48 h after
96 h after
Series HI
(Day 14 denervation)*
Maintenance
48 h before
At time of
Restimulation
24 h after
48 h after
Limbs with midcone stage outgrowths or better 35 days postamputation were counted as regenerates.
•Time of denervation after amputation.
Ill data also show that longer delays between denervation and DRG grafting resulted in an increased
frequency of very small spike regenerates in both
denervated controls and DRG-implanted limbs. This
suggests that DRG implants stimulated the growth of
regenerates which, at the time of implantation, contained a relatively large blastema cell population
bordering on nerve-independence. Moreover, the
maintenance strategy was more effective in stimulating substantial regenerative outgrowth thus
avoiding the minimal nerve-independent responses
observed on denervated control limbs and some
DRG-implanted limbs when the restimulation strategy was used.
Xenografts to denervated axolotl limbs
Controls
In Series IV and V, limbs that remained innervated
throughout the experiment had blastema-stage regenerates 7 days postamputation and three- to four-digitstage regenerates 12 to 14 days postamputation. Of
the limbs allowed to become reinnervated following a
single denervation (n = 14), three progressed to the
midcone stage within the 16-day experimental period;
had these reinnervated limbs been fixed later, it is
likely that a larger percentage would also have
regenerated (Petrosky et al. 1980). Redenervated,
nonimplanted and liver- or muscle-implanted control
limbs did not regenerate (Table 3) but resorbed to the
level of the proximal radius-ulna or distal humerus.
Maintenance and restimulation strategies
In Series IV, DRG xenografted at the time of
denervation (24 h after amputation) to maintain
neurotrophic stimulation, promoted blastema formation and regenerate outgrowth on 50% of the
implanted limbs (Table 4). DRG implanted 96 h after
denervation, to maintain cell cycling and resupply
neurotrophic activity, promoted regeneration on
90% of the implanted limbs. When DRG were
grafted 8 days after denervation, to restimulate cell
cycling and resupply neurotrophic activity, 53 % of
the limbs were stimulated to regenerate. At each time
interval, implanted DRG promoted a significant
stimulation of regeneration (P< 0-001) compared to
denervated limbs (with or without control tissue
implants). Limbs implanted with DRG on day 4
regenerated at a significantly greater frequency
(P<004) than limbs implanted 8 days after nerve
withdrawal. Comparisons between other implant
groups did not show statistically significant differences (P> 0-05).
The morphological stage attained varied with the
timing of DRG implantation. Table 4 data show that
the most advanced regenerates (palette and notch
stages) were observed among limbs implanted 4 days
DRG stimulation of denervated urodele forelimb regeneration
Table 3. Regenerative responses observed in larval
axolotl control limbs
Implant group
Number
regenerating (%)
n
Series IV
(Day 1 denervation)*
Innervated
Denervated
Reinnervated
liver
At the time of
4 days after
8 days after
Muscle
At the time of
4 days after
8 days after
4
10
14
4(100)
0
3(21)
4
4
4
0
0
0
4
4
4
0
0
0
4
4
4(100)
0
Series V
(Day 4 denervation)*
Innervated
Denervated
* Time of denervation after amputation.
after denervation. While two palette regenerates
were observed on limbs implanted 8 days after
denervation, midcone stage regenerates were most
common and with concomitant denervation and
DRG grafting only midcone stage regenerates were
observed.
In Series V, regeneration occurred in five of six
limbs implanted with a pair of DRG at the time of
denervation (4 days after amputation). All five limbs
reached at least the cone stage of regeneration and
three of these progressed to the three-digit stage
(Table 4); one limb did not retain the graft and
resorbed. All six limbs that received DRG grafts 4
days after denervation regenerated to the midcone
stage (Table 4).
Histology
All tissues expected in a mature limb (muscle, cartilage and connective tissue) were present in DRGstimulated regenerates; the DRG remained at or near
the original plane of amputation (Figs 4, 5). In
implanted adult newt limbs the homoplastically
grafted DRG were invariably obscured by an accumulation of cells, probably the result of an immunological rejection of the graft; as a consequence
neuronal cell numbers were substantially reduced
(Fig. 4C,D).
Fig. 6 is representative of redenervated nonimplanted axolotl limbs and redenervated limbs with
liver or muscle implants; all three of these control
limb groups exhibited only pseudoblastemata typical
of resorbing denervated larval limbs (Schotte" &
Butler, 1941; Olsen & Tassava, 1984). In axolotl limbs
that were stimulated to regenerate by DRG implants,
large numbers of blastema cells were present at the
distal end of the limb (Fig. 5). The cellular accumulation obscuring DRG homografts in adult newt limbs
was not observed in association with DRG xenografts
in axolotl limbs (compare Figs 4C,D and 5C,D). With
respect to the original plane of amputation, the
position of the DRG in the distal stump varied
somewhat from limb to limb (compare Figs 5 and 7),
but whether a correlation existed between the final
location of the DRG and the extent/frequency of
regeneration could not be ascertained.
In newt limbs nerve fibre growth associated with
the implanted DRG was not observed. In axolotl
Table 4. Stimulation of larval axolotl forelimb regeneration by DRG xenografts
Implant group
Nonregenerate
Series IV
(Day 1 denervation)*
Maintenance
At time of
Maintenance and/or
Restimulation
4 days after
Restimulation
8 days after
8
4
10
1
19
9
6
1
6
0
Series V
(Day 4 denervation)*
Maintenance
At time of
Restimulation
4 days after
'Time of denervation after amputation.
179
Spike
Midcone
Late
cone
Palette
3-4
digit
180
B. L. Tomlinson and R. A. Tassava
forelimb regenerates, staining revealed nerve fibre
outgrowth from DRG in 82 % (28/34) of the limbs. In
six regenerates nerve fibre growth from the DRG was
not observed (Fig. 8B), in eighteen regenerates the
ganglia produced only a few nerve fibres (Fig. 8C)
and in ten cases a large number of nerve fibres
coursed through the regenerate and directly into the
apical epidermis (Fig. 8D). In nonregenerating
DRG-implanted limbs, nerve fibre growth from the
ganglia was observed in only two of fifteen cases and
in both instances the fibres were restricted to the
margins of the DRG implants (similar to that shown
in Fig. 8B).
Discussion
DRG grafts provide the neurotrophic stimulation
essential for forelimb regeneration in both adult
DRG stimulation of denervated urodele forelimb regeneration
newts and larval axolotls. Excision and immediate
autografting of brachial ganglia (Kamrin & Singer,
1959) has been shown to stimulate continued regeneration of the adult newt limb. While this procedure
eliminates the sensory supply, without redenervation
it does not prevent motor nerves from regenerating
into the limb and supplementing the neurotrophic
stimulation offered by the grafted DRG. The present
procedure of homografting DRG and repeatedly
transecting the host brachial plexus demonstrated
that DRG did stimulate forelimb regeneration independent of the host nerve supply. Furthermore,
pituitary gland implants were ineffective stimulators
of blastema outgrowth arguing against the view that
pituitary secretions may be functionally identical to
neurotrophic factors (Wallace, 1981).
In adult newts, the efficacy of DRG stimulation of
regeneration is related to the number of blastema
cells present at the time of denervation and the timing
of DRG implantation with respect to denervation.
With each increase in time prior to denervation and
DRG grafting (1, 10 and 14 days), larger percentages
(0, 34 and 57 %) of adult newt limbs regenerated,
thus suggesting a correlation between DRG stimulation and the known increase in blastema cell numbers through this time period (Chalkley, 1954). As the
interval between amputation and denervation was
extended (to 14 days) in denervated control limbs,
the development of small nerve-independent spikes
(Singer & Craven, 1948) was sometimes observed,
presumably due to the accumulation of a stable
population of blastema cells. While denervation on
day 10 resulted in a total inhibition of control limb
regeneration, it is significant that even with the
smaller population of blastema cells, implanted DRG
Fig. 4. Denervated adult newt forelimbs stimulated to
regenerate following the implantation of two dorsal root
ganglia (G). (A) A late-cone-stage regenerate formed
following the implantation of ganglia 1 day after
denervation (Series II). The bar represents 250 fjm.
(B) This four-digit-stage regenerate developed following
the implantation of ganglia at the time of denervation
(Series III). While two digits are present in this section,
the remaining portions of the other digits are observed
only in adjacent sections. The bar represents 250fan. (C)
An enlargement of the box in (B) to show a portion of
the region containing a representative implanted
ganglion. Note the dense accumulation cf cells, which
obscure the ganglia, and the reduced number of neuronal
cell bodies (arrows). The bar represents 50 fan. (D) A
further enlargement of the ganglion in C to show the few
neurones (arrows) found in the limb. The bar represents
20 fan. Compare the appearance of the ganglia implants
in Fig. 4 to those in Fig. 5. In A and B the approximate
plane of amputation (midhumerus) is marked by a dashed
line. In C and D regeneration is directed toward the top
of the micrograph. Haematoxylin and eosin staining.
181
nevertheless stimulated regeneration at this early
nerve-dependent stage. Furthermore, the efficacy of
stimulation by ganglia on day 14 was demonstrated by
the increased size and morphological complexity of
the regenerates on experimental limbs compared to
the nerve-independent spikes observed on denervated control limbs.
The protocol of reducing residual neurotrophic
activity by denervating for some period of time and
subsequently implanting DRG to restimulate regeneration has not previously been investigated in vivo.
Kamrin & Singer (1959) only used the maintenance
strategy in their study which demonstrated that
autoplastically grafted DRG promoted outgrowth of
regenerates on newt forelimbs. Similarly, the maintenance strategy using cocultured DRG in vitro
(whereby innervated regenerates were excised and
immediately cocultured with DRG) has been shown
to stimulate macromolecular synthesis (VethamanyGlobus et al. 1978) and mitosis (Globus &
Vethamany-Globus, 1977; Tomlinson et al. 1981).
However, based on data from a combined in vivo and
in vitro study, Tomlinson et al. (1984) suggested that
blastema cells may become refractory to neurotrophic stimulation following denervation. Regenerates that had been denervated for 48 h in vivo did not
subsequently respond to DRG stimulation when
explanted in vitro (i.e. restimulation). Other studies
using extracts or purified molecules have only
employed the restimulation strategy to investigate the
neurotrophic stimulation of regeneration. Infusions
of purified molecules (Mescher & Gospodarowicz,
1979; Carlone & Rathbone, 1985) or nerve extracts
(Lebowitz & Singer, 1970; Singer et al. 1976; Jabaily
& Singer, 1977; Carlone & Foret, 1979) 48 h after
denervation stimulated only limited increases in biological activity and did not result in regenerate
outgrowth (Wallace, 1981).
Results of the present investigation in which maintenance and restimulation strategies were compared
are consistent with the view (Tomlinson et al. 1984)
that, soon after denervation, adult newt blastemata
lose responsiveness to restimulation by nerves or
nerve extracts. As the length of the denervation time
was increased prior to ganglia implantation, successively fewer newt limbs regenerated. DRG implanted
before, or concomitant with, denervation stimulated
regeneration, but their effectiveness was reduced
after only a 24 h denervation time suggesting relatively rapid changes within the blastema. It is interesting to note that some parameters indicative of cell
cycle activity (DNA synthesis, Singer & Caston, 1972;
mitosis, Tomlinson etal. 1984) remain elevated for up
to 48 h after denervation. In this regard, we can now
investigate the DRG stimulation of regeneration of
182
B. L. Tomlinson and R. A. Tassava
Fig. 5. Denervated axolotl forelimbs stimulated to regenerate following the implantation of a single dorsal root ganglion
(G). (A) An early three-digit regenerate (digits are numbered 1, 2 and 3) formed following the implantation of a
ganglion 1 day after denervation (Series IV). The approximate location of the ganglion (G; not seen in this section) is
just distal to the plane of amputation (midradius-ulna; dashed line). The bar represents 200/an. (B) A ganglion (G)
implanted 4 days after denervation (Series IV) stimulated the growth of this palette-stage regenerate. Note the large
accumulation of blastema cells distal to the ganglion which is near the plane of amputation (dashed line). The bar
represents 175 /an. (C) An enlargement of half of the ganglion in B showing the abundant neuronal cell bodies, some of
which are indicated by arrows. The bar represents 100/mi. (D) A further enlargement of the ganglion in C to show the
abundant neurones (arrows) found in these limbs. The bar represents 50/an. In C and D regeneration is directed toward
the top of the micrograph. Haematoxylin and eosin staining.
DRG stimulation of denervated urodele forelimb regeneration
Fig. 6. Histological section through a denervated control
limb implanted with newt flank muscle. Note the small
accumulation of cells indicative of a pseudoblastema
(pb). Due to resorption following denervation, the
original plane of amputation was distal to, but
approximately parallel with, the dashed line. Evidence of
continued resorption is seen as distal portions of the
humerus (h) undergo histolysis. Bar represents 200 ^m.
Haematoxylin and eosin staining.
denervated adult newt limbs in relation to the size
and proliferative status of the blastema.
The present study provides the first evidence that
denervated larval axolotl limbs are also capable of
regenerating in response to implanted DRG. Both
maintenance and restimulation strategies were effective in promoting the regeneration of denervated
axolotl limbs at a relatively high frequency apparently
independent of the size of the blastema and timing of
DRG implantation with respect to denervation. It is
of interest to consider the timing of DRG implantation relative to amputation in axolotls with regard
to the development of an in vivo bioassay. Between 4
and 7 days following simultaneous amputation and
denervation there is a limited and transient burst of
183
cell cycle activity (Maden, 1978) which is subsequently suppressed (Olsen et al. 1984; Barger &
Tassava, 1985). Therefore DRG implanted concomitant with, or 4 days after, denervation (Series IV, 1
and 5 days after amputation, respectively) stimulated
regeneration by maintaining cell cycle activity. In
addition, DRG grafted concomitant with denervation
4 days after amputation (Series V) also maintained
cycling activity and stimulated regeneration. By 8
days after amputation residual cell cycle activity was
suppressed in those limbs that had been denervated
for 4 (Series V) or 8 (Series IV) days (Tassava &
McCullough, 1978; Maden, 1978) and therefore DRG
implants presumably restimulated cell cycling to promote regeneration. Grafted dorsal root ganglia therefore are a complete source of factors necessary to
restimulate regeneration-specific events in denervated axolotl limbs. The present DRG in vivo system
can now be used to investigate short-term cell cycle
events by both the maintenance and restimulation
strategies and, in addition, it provides the opportunity to examine sustained growth and morphogenesis.
The ability of newt DRG to stimulate regeneration
in denervated axolotl limbs reflects the fact that
neurotrophic factors lack species specificity (Kamrin
& Singer, 1959). Why then were differences in
response to DRG observed between newts and axolotls? How can it be explained that DRG stimulated a
higher frequency of regeneration in axolotl limbs
after prolonged denervation and why was nerve fibre
growth from the grafted ganglia more commonly
observed in axolotl regenerates? At least part of the
reason may involve the faster rate of regeneration
(Tomlinson et al. 1985) and apparent immunological
tolerance of the graft in axolotls. Evidence for immunological graft rejection was observed in adult newt
limbs but not in larval axolotl limbs. Since both newts
and axolotls are known to be capable of rejecting
7A
Fig. 7. (A) A nonregenerating denervated limb that contained an implanted ganglion. Note the large accumulation of
cartilage (c) at the end of the limb. (B) Another section of the limb in A which shows that the position of the ganglion
(G) is a considerable distance proximal to the original plane of amputation (dashed line). The bar represents 200 fan.
Haematoxylin and eosin staining.
184
B. L. Tomlinson and R. A. Tassava
Fig. 8. (A) A typical cone-stage regenerate on a denervated axolotl limb stimulated by an implanted dorsal root
ganglion (G). Note the distal accumulation of blastema cells. The bar represents 250,um. Haematoxylin and eosin
staining. The remaining photomicrographs are enlargements of silver-stained sections showing nerve fibre growth
associated with the ganglia (which is located on the left of each micrograph). (B) This DRG supported the regeneration
of the axolotl limb shown in Fig. 5B. Fibre growth was limited to within the margins of the ganglion and few fibres (if
any) could be seen within the blastema. (C) The ganglion shown here is the same one shown in Fig. 8A. Note the
numerous fibres that extend beyond the margins of the ganglion toward the regenerate. Nerve fibres were seen scattered
throughout this regenerate. (D) An example of abundant nerve fibre growth seen from an implanted DRG. This dense
accumulation of fibres eminating from the DRG coursed through the associated cone-stage regenerate (not shown). For
the purpose of orientation the direction of regeneration in each of B, C and D is towards the right of the micrograph
and the bar in each of these figures represents 100 fim.
tissue grafts (reviewed by Wallace, 1981), the lack of
such a response in the larval axolotl limb may be the
consequence of the faster. completion of regeneration, resulting in fixation for histology prior to the
onset of a graft-rejection response.
Regardless of immunological complications,
grafted DRG stimulated some regeneration in each
species using both maintenance and restimulation
strategies; however, while DRG were equally effective in both maintenance and restimulation strategies
in axolotls, the restimulation strategy was considerably less effective than the maintenance strategy in
stimulating the denervated adult newt limb to regenerate. Differing responses to DRG implants may also
be related in part to different intrinsic tissue responses to neurotrophic stimulation. For example,
the reinnervation of a previously denervated newt
limb is not sufficient to stimulate regeneration (the
limb must also be reamputated; Salley & Tassava,
1981), but in axolotls reinnervation is sufficient to
DRG stimulation of denervated wodele forelimb regeneration
stimulate regeneration (Schott6 & Butler, 1949;
Petrosky et al. 1980).
Utilization of the larval system as a reproducible
regeneration-specific bioassay is a possibility worthy
of further exploration. Outgrowth of a regenerate on
a denervated limb, stimulated by the appropriate
neurotrophic treatment, is a rigorous but ultimately
essential assay for neurotrophic activity. Since the
present studies show that DRG implants do stimulate
blastema formation and growth, future experiments
can be designed to examine the macromolecular
and/or cell cycle events that occur shortly after
implantation of DRG. Once the specific events
influenced by grafted DRG have been defined it will
be possible to determine if various DRG extracts
and/or putative neurotrophic molecules influence the
parameters in an identical manner.
This research was supported by NSF grant PCM 8315428
and USPHS grant NS-10165. Critical review of the manuscript by David J. Goldhamer is appreciated. We would also
like to thank Donna E. Tomlinson for technical assistance
and preparation of the manuscript.
References
P. M. & TASSAVA, R. A. (1985). Establishment
of a regeneration-specific in vivo bioassay for
neurotrophic activity in denervated Ambystoma
forelimbs. Experientia 41, 1405-1407.
BUTLER, E. G. & SCHOTTE, O. E. (1949). Effects of
delayed denervation on regenerative activity in limbs of
urodele larvae. J. exp. Zool. 112, 361-392.
CARLONE, R. L. & FORET, J. E. (1979). Stimulation of
mitosis in cultured limb blastemata of the newt,
Notophthalmus viridescens. J. exp. Zool. 210, 245-252.
CARLONE, R. L. & RATHBONE, M. P. (1985). Partial
purification of a low molecular-weight growth factor
from chicken brain. /. cell. Physiol. Y2A, 467-473.
CHALKLEY, D. T. (1954). A quantitative histological
analysis of forelimb regeneration in Triturus viridescens.
J. Morph. 94, 21-70.
DRESDEN, M. H. (1969). Denervation effects on newt
limb regeneration: DNA, RNA, and protein synthesis.
Devi Biol. 19, 311-320.
GLOBUS, M. (1978). Neurotrophic contribution to a
proposed tripartite control of the mitotic cycle in the
regeneration blastema of the newt, Notophthalmus
viridescens. Am. Zool. 18, 855-868.
GLOBUS, M. & VETHAMANY-GLOBUS, S. (1977). Transfilter
mitogenic effect of dorsal root ganglia on cultured
regeneration blastemata, in the newt, Notophthalmus
viridescens. Devi Biol. 56, 316-328.
BARGER,
GLOBUS, M., VETHAMANY-GLOBUS, S., KESIK, A. &
MILTON, G. (1983). Roles of neural peptide substance
P
and calcium in blastema cell proliferation in the newt
Notophthalmus viridescens. Limb Development and
Regeneration, part A (ed. J. F. Fallon & A. I. Caplan),
pp. 513-524. New York: Alan R. Liss Inc.
185
ITEN, L. E. & BRYANT, S. V. (1973). Forelimb
regeneration from different levels of amputation in the
newt, Notophthalmus viridescens: Length, rate and
stages. Wilhelm Roux' Arch. EntwMech. Org. 173,
263-282.
JABAILY, J. A. & SINGER, M. (1977). Neurotrophic
stimulation of DNA synthesis in the regenerating
forelimb of the newt, Triturus. J. exp. Zool. 199,
251-256.
KAMRIN, A. A. & SINGER, M. (1959). The growth
influence of spinal ganglia implanted into the
denervated forelimb regenerate of the newt, Triturus.
J. Morph. 104, 415-439.
LEBOWITZ, P. & SINGER, M. (1970). Neurotrophic control
of protein synthesis in the regenerating limb of the
newt, Triturus. Nature, Lond. 225, 824-827.
MADEN, M. (1978). Neurotrophic control of the cell cycle
during amphibian limb regeneration. /. Embryol. exp.
Morph. 48, 169-175.
MESCHER, A. L. & GOSPODAROWICZ, D. (1979). Mitogenic
effect of a growth factor derived from myelin in
denervated regenerates of newt forelimbs. /. exp. Zool.
207, 497-503.
MESCHER, A. L. & LOH, J. J. (1981). Newt forelimb
regeneration blastemas in vitro: cellular response to
explanation and effects of various growth promoting
substances. /. exp. Zool. 216, 235-245.
MESCHER, A. L. & MUNIAM, S. I. (1984). Trophic effect
of transferrin on amphibian limb regeneration
blastemas. /. exp. Zool. 230, 485-490.
MESCHER, A. L. & TASSAVA, R. A. (1975). Denervation
effects on DNA replication and mitosis during the
initiation of limb regeneration in adult newts. Devi
Biol. 44, 187-197.
OLSEN, C. L. & TASSAVA, R. A. (1984). Cell cycle and
histological effects of reinnervation in denervated
forelimb stumps of larval Ambystoma. J. exp. Zool. 229,
247-258.
OLSEN, C. L., BAKGER, P. M. & TASSAVA, R. A. (1984).
Rescue of blocked cells by re-innervation in denervated
forelimb stumps of larval Ambystoma. Devi Biol. 106,
399-405.
PETROSKY, N. S., TASSAVA, R. A. & OLSEN, C. L. (1980).
Cellular events in denervated limb stumps of
Ambystoma larvae during re-innervation and
subsequent regeneration. Experientia 36, 601-603.
SALLEY, J. D. & TASSAVA, R. A. (1981). Responses of
denervated adult newt limb stumps to reinnervation
and reinjury. /. exp. Zool. 215, 183-189.
SAMUEL, E. P. (1953). Towards controllable silver
staining. Anat. Rec. 116, 511-517.
SCHOTTE, O. E. & BUTLER, E. G. (1941). Morphological
effects of denervation and amputation of limbs in
urodele larvae. /. exp. Zool. 87, 279-322.
SINGER, M. (1952). The influence of the nerve in
regeneration of the amphibian extremity. Q. Rev. Biol.
27, 169-200.
SINGER, M. & CASTON, J. D. (1972). Neurotrophic
dependence of macromolecular synthesis in the early
limb regenerate of the newt, Triturus. J. Embryol. exp.
Morph. 28, 1-11.
186
B. L. Tomlinson and R. A. Tassava
M. & CRAVEN, L. (1948). The growth and
morphogenesis of the regenerating forelimb of adult
Triturus following denervation at various stages of
development. J. exp. Zool. 108, 279-308.
SINGER,
SINGER, M., MAIER, C. E. & MCNUTT, W. S. (1976).
Neurotrophic activity of brain extracts in forelimb
regeneration of the urodele, Triturus. J. exp. Zool. 196,
131-150.
TASSAVA, R. A. (1969). Hormonal and nutritional
requirements for limb regeneration and survival of
adult newts. /. exp. Zool. 170, 33-54.
TASSAVA, R. A. & MESCHER, A. L. (1975). The roles of
injury, nerves and wound epidermis during the
initiation of amphibian limb regeneration: an
hypothesis. Differentiation 4, 23-24.
TASSAVA, R. A. & MCCULLOUGH, W. D. (1978). Neural
influence on cell cycle events during limb regeneration.
Am. Zool. 18, 729-740.
TOMLINSON, B. L., GOLDHAMER, D. J., BARGER, P. M. &
TASSAVA, R. A. (1985). Punctuated cell cycling in the
regeneration blastema of urodele amphibians: an
hypothesis. Differentiation 28, 195-199.
TOMLINSON, B. L., GLOBUS, M. & VETHAMANY-GLOBUS,
S. (1981). Promotion of mitosis in cultured newt limb
regenerates by a diffusible nerve factor. In Vitro 17,
167-172.
TOMLINSON, B. L., GLOBUS, M. & VETHAMANY-GLOBUS,
S. (1984). Blastema cell cycle in vitro and attempted
restimulation of blastema cell cycling in denervated
blastemata of the adult newt, Notophthalmus
viridescens. J. exp. Zool. 232, 249-258.
VETHAMANY-GLOBUS, S., GLOBUS, M. & TOMLINSON, B.
L. (1978). Neural and hormonal stimulation of DNA
and protein synthesis in cultured regeneration
blastemata in the newt, Notophthalmus viridescens. Devi
Biol. 65, 183-192.
VETHAMANY-GLOBUS, S., GLOBUS, M., DARCH, A.,
MILTON, G. & TOMLINSON, B. L. (1984). In vitro effects
of insulin on macromolecular synthesis events in newt
limb regeneration blastemata. J. exp. Zool. 231, 63-74.
WALLACE, H. (1981). Vertebrate Limb Regeneration. New
York: John Wiley and Sons.
{Accepted 13 October 1986)