THE JOURNAL OF COMPARATIVE NEUROLOGY 300:37&385 (1990)

Collateral Sproutingof the Central Temninalsof Cutaneous PrimaryMerent Neurons in the Rat Spinal Cord:Pattern, Morphology,and Influence of Targets MARIA FITZGERALD, CLIFFORD J. WOOLF, AND PETER SHORTLAND Department of Anatomy and Developmental Neurobiology, University College London, London WClE 6BT, England

ABSTRACT The capacity of the central terminals of primary afferents to sprout into denervated areas of neonatal spinal cord and the morphology of any novel terminals has been investigated. In rats which had undergone sciatic nerve section on the day of birth, 12 of 18 physiologically characterized intact saphenous hair follicle afferents (HFAs) were labelled intra-axonally with horseradish peroxidase (HRP) were shown to sprout up to 2,000 Fm into the deafferented sciatic terminal field. The morphology of these sprouts depended on which area of the sciatic nerve territory was invaded by the afferent sprouts. Six HFAs sprouted into areas normally innervated by glabrous skin afferents and the morphology of the collateral sprouts in this region resembled that of rapidly adapting (RA) afferents. The other six saphenous HFAs had sprouted into sciatic "hairy" skin areas and the morphology of these sprouts, although abnormal, was flame shaped. In rats whose sural, saphenous, and superficial peroneal nerves were cut at birth, 4 of 7 single HRP labelled RA afferents had central terminals that had sprouted into regions of cord normally devoted to "hairy" input. These showed clear signs of HFA morphology despite their peripheral receptive fields remaining in the glabrous skin. The results show collateral sprouting of single cutaneous sensory afferent axons into adjacent inappropriate central target regions following neonatal deafferentation. Such plasticity may provide some compensation following neonatal injury. The morphology of the sprouted terminals is appropriate to the new target area rather than to its functional class and is also independent of the peripheral receptive field location providing an example of central rather than peripheral control over afferent growth patterns. Key words: primary afferent, sprouting,horseradishperoxidase,spinal cord, dorsal horn

Collateral axonal sprouting, or the triggering of intact axons to form new terminal branches in order to establish connections with vacant synaptic sites, has been observed in a number of areas of the mammalian central nervous system following partial denervation. The phenomenon is most commonly observed in the developing nervous system (Kalil and Schneider, '75; Stanfield and Cowan, '76; Fitzgerald, '85; Rhoades et al., '89) but also occurs in the adult (Raisman and Field, '73; Tsukahara, '81; Cotman et al., '81; Molander et al., '88; Lamotte et al., '89). Sprouting does not appear to occur however in all circumstances where there is a deafferentation (Guillery, '72; White and Nolan, '74; Kerr, '75; Rustioni and Molenaar, '75; Rodin et al., '83; Rodin and Kruger, '84; Stelzner and Devor, '84). The factors that initiate neurite outgrowth in an uninjured neuron and then guide the growing neurites to a new

o 1990 WILEY-LISS, INC.

target are not known (but see Diamond et al., '871, nor is it known what determines the morphology of the sprouts; the intrinsic properties of the sprouting neuron or the target it innervates. In order to help address this latter problem we have investigated the central collateral sprouting of single physiologically identified primary afferent neurons in the rat spinal cord by filling individual axons with horseradish peroxidase (HRP) following neonatal deafferentation. The central terminals of cutaneous primary afferents are somatopically highly ordered (Swett and Woolf, '85; Woolf and Fitzgerald, '86). Section of a peripheral nerve on the day of birth results in a substantial cell death of the axotomized neurons (Yip et al., '84; Bondok and Sansone, '841,producing a central denervation. This results in an Accepted July 19,1990,

CENTRAL SPROUTING OF PRIMARY AFFERENTS

371

\

\

E

I

/

/

Fig. 1. Camera lucida reconstructions of 6 adjacent terminal arborizations from rostra1 to caudal (A-F) of a control saphenous nerve HFA whose receptive field was on the medial leg. Scale bar, 250 pm.

expansion, which has been revealed by bulk labelling techniques, of the central terminals of neighboring intact nerves into the dennervated site in the dorsal horn (Fitzgerald, '85). The expansion is apparantly due to genuine new collateral growth and not failure of exuberant terminals to withdraw since such exuberance is not displayed by developing primary afferents in the spinal cord (Smith, 1983; Fitzgerald, '85). A study of the consequences of a neonatal denervation on single intact neighboring afferent neurons

allows therefore the incidence, extent, and terminal morphology of collateral axonal sprouts to be analyzed.

MATERlALSANDME'IXODS Neonatalnervesections Rat pups were removed from their litter at birth and anaesthetized by cooling to 5°C. In one group the sciatic

M. FITZGERALD ET AL.

372

Snow, ’84) and rat (Woolf, ’87; Shortland et al., ’89). The HFAs had a stem axon that ascended in the dorsal columns, giving off up to 10 collateral axons into the dorsal horn. The majority of these collaterals terminated after a characteristic dorsal loop of the collateral axon as complex “flameshaped” arborizations dense with synaptic boutons (Fig. 1) which overlapped rostrocaudally to form a continuous, narrow longitudinal sheet of terminal boutons. At the rostral and caudal ends of these sheets lay simple arbors with fewer branch points and sparse terminal boutons and beyond them blind collaterals which had no boutons at all (Woolf, ’87; Shortland et al., ’89).In each of the ten HFAs IntracelIularreCordingandlabe~ the mediolateral, dorsoventral, and rostrocaudal dimenThe rats were anesthetized with urethane (1.5g kg-’, i.p.) sions of the complex arbors; the intercollateral spacing; and and then decerebrated by aspiration of all cranial contents the distance between the most rostral and caudal collateral rostral to the midbrain. Following this they were paralyzed were measured (Table 1). with gallamine (Flaxedil) and artificially ventilated. The Somatotopic location of terminals. The location of vertebral column was clamped at thoracic and sacral levels the central terminal fields of the 10 control saphenous and the hips fixed with plaster of paris to hip bars for HFAs in the dorsal horn of the spinal cord were exactly as further stabilization. The lumbar spinal cord was exposed expected from previous transganglionic studies of the sapheand after removal of the dura, a small (0.5 cm) supporting nous nerve’s central terminal field (Swett and Woolf, ’85; plate slipped underneath it such that the cord was held just Woolf and Fitzgerald, ’86; Molander and Grant, ’86) and above the floor of the vertebral canal (Woolf and King, ’87). from a study of the somatotopic distribution of individual The cord was then covered with 1%agar and a small area HFAs (Shortland et al., ’89). The central terminal fields exposed and covered with warm mineral oil for recording. were all restricted to a zone in the medial part of the caudal Intracellular recordings were made from single afferent L2, the L3, and the rostral L4 segments of the lumbar fibres in the dorsal root entry zone of the lumbar cord using enlargement (Fig. 2A). This saphenous terminal zone is thin-walled microelectrodes filled with 5%HRP in TrisiKCl known from previous studies to be surrounded laterally and buffer (pH 7.7, impedances 20-40 Ma). Receptive fields caudally by sciatic nerve terminals from the hairy skin of were carefully mapped and characterized using a fine camel the lateral (sural) and dorsal (superficial peroneal) surfaces hair brush, von Frey hairs, and a blunt probe, and conduc- of the leg and foot (marked H in Fig. 2A) and medially by tion velocities estimated from electrical stimulation (5 mA, sciatic nerve terminals from the glabrous skin on the sole of 500 pS, 1 Hz) using skin pin electrodes. HRP was then the foot (tibial area, marked Gin Fig. 2A) (Swett and Woolf, injected into the fibre using 150 ms depolarizing pulses ’85; Molander and Grant, ’86; Shortland et al., ’89). The every 200 ms for 2-10 minutes. A minimum of 2 hours later caudal border of the saphenous terminal field with the the rat was perfused with cold saline followed by 1.25% sciatic territory lies at the mid L4 segment. The definition glutaraldehyde, 1.0%paraformaldehyde in 0.1 M phosphate of the saphenous terminal field used here is the area buffer, pH 7.4, 4°C. Lumbar segments were identified by occupied by both complex and simple collateral arbors (ie tracing the entry of dorsal roots and marked with insect bouton containing arbors). Blind ending collaterals do lie pins. The cords were removed, stored overnight in 20% outside this field (Shortland et al., ’89); and their most sucrose in 0.1 M phosphate buffer at 4°C. Serial transverse caudal extent, which is beyond mid L4, is indicated on sections (50 km) were cut and stained for the HRP reaction Figure 2A for the present sample. In L4 the saphenous product using the Hanker-Yatesmethod (Hanker et al., ’77) arbors never normally extend to the medial edge of the as described previously (Shortland et al., ’89). The termi- dorsal horn (Fig. 2A). nals of individual afferents were reconstructed from all the sections using a camera lucida.

nerve (comprising the sural, superficial peroneal, and tibial nerves) was exposed in the upper thigh, ligated, and cut. These animals will be referred to as “sciatic sectioned” rats. In the other group, the sural, superficial peroneal, and saphenous nerves were ligated and cut, leaving the tibial nerve intact. These animals will be referred to as “tibial intact” rats. The muscle and skin wounds were sutured, the pups allowed to recover body temperature, and then returned to their litters and allowed to grow into adults (weight 200-250 g).

TABLE 1. Dimensions of Hair Follicle Afferent Terminals’ Sciaticcut

RESULTS The morphology and somatotopic organization of the central terminals of hair follicle afferents with receptive fields in the saphenous nerve territory, and of rapidly adapting afferents with receptive fields on the glabrous skin of the toes (tibial nerve territory), have been studied in control animals and in animals with either sciatic nerve section on the day of birth, “saphenous intact animals,” or where the sural, saphenous, and common peroneal nerves were sectioned on the day of birth, “tibial intact animals.”

Control saphenoushair follicle af€erents Morphology. Ten hair follicle afferents (HFAs) from control animals (all peripheral nerves intact) were found to have the morphological and receptive field characteristics typical of this class of afferent that have been previously described both in the cat (Brown et al., ’77; Meyers and

Control

Rostral to L3iz4 border

Caudal to L3L4 border

10

6

12

8.6 t 1.0 4 3 t 0.3 1.7 ? 0.6 2.6 f 0.5 342 26

7.7 ? 1.3 3.5 2 0.5 2.3 5 0.8 2.0 t 0.9 376 t 36

8.8 :+ 0.9 3 8 2 0.7 3.0 t 0.6 2.0 I 0.3 381 :+ 27

2.4 2 0.3

3.0 t 0.3

87 i 4 55 t 5

116 t 15

146 ? 13 88 t 8

201 c 14 113 t 15

232 -i 30 8 6 2 11

229 c 19 135 ? 15

269 ? 16 129 t 13

253 t 18 129 i 15

238 t 14 163 t 17

N Mean No. of collateralv Total Complex Simple Blind-ending Intercollated distance, p,m Total rostrocaudal collateral distance, mm Arbor dimensions, pm Mdolateral Complex Sunple Dorsoventral Complex Sunple Rostrccaudd Complex Simple

‘x t SEM

2.8

2

0.3

50 i 7

CENTRAL SPROUTING OF PRIMARY AFFERENTS

M

I

373

M

1

I

RECEPTIVE FJELD

J

:

I - -

i

500 pm

H

- - -

- - -

i

H

- - -

- LWL6

A Fig. 2. Plan views through lamina I11 of the dorsal horn showing the terminal distribution of bouton containing collaterals of saphenous nerve HFAs. A Control saphenous HFAs. B: Saphenous HFAs in animals whose sciatic nerve was sectioned at birth. M represents the medial border and L the lateral border of the dorsal horn, and L2 to L6 represent the lumbar segments divided by horizontal dashed lines. The segment lengths are means from all the animals used and the afferent terminals have been plotted by taking L3/L4 as an absolute boundary. The mediolateral width of the collateral arborizations have been taken as a percentage of the width of the dorsal horn. Note that in B the L3-L5 dorsal horn is 25% narrower as a result of neonatal deafTerenta-

tion. The curved outlines within the dorsal horn are the borders formed by the terminal fields of the total sample of filled saphenous HFAs from each group (n = 10 for A; n = 18 for B). Examples of the collateral arborizations of individual HFAs are shown as shaded areas and their cutaneous receptive fields on the medial leg and toes can be seen on the right hand column of the figure. The arrows labelled 1-6 on the medial side of the plans mark the position of the most caudal blind-ending collateral of the individual illustrated afferents, numbered in the right-hand panel. The letters G and H stand for glabrous and hairy, respectively, to illustrate the areas of cord normally entirely devoted to either glabrous or hairy skin inputs.

M. FITZGERALD ET AL.

374

\

Fig. 3. Camera lucida reconstructions of 6 adjacent terminal arborizations from rostral to caudal (A-F) of a toe glabrous skin rapidly adapting mechanoceptor from the tibial nerve. Scale bar, 250 pm.

Control tibial glabrousskin afferents Morphology. Eight rapidly adapting mechanoreceptors innervating the skin of the toes were recovered in control animals. As previously reported (Woolf, ’87), the

stem axons of these afferents travel in the dorsal columns giving off up to 12 largely non-overlapping collaterals generating arbors of the complex, simple and blind-ending type in the dorsal horn (Fig. 3). The dimensions of these

Fig. 4. Plan view through lamina IV of the dorsal horn showing the terminal distribution of bouton-containing collaterals of tibial nerve toe glabrous skin rapidly adapting mechanoceptors. A: Control tibial toe mechanoceptors (n = 8). B: Tibia1 toe mechanoceptors from animals where only the tibial nerve was left intact as a neonate (n = 7). The rostral shift between A and B is unlikely to be significant since the field

remains within the tibial area. The medial shift is significant. Shrinkage of the dorsal horn of the lumbar segments due to nerve section is 16-20%. The numbered arrows represent the position of the most rostral collateral of an individual afferent. For detailed explanation of plan view, see Figure 2.

CENTRAL SPROUTING OF PRIMARY AFFERENTS

375

w

0 W

a

I

I

I

1

M. FITZGERALD ET AIL.

376 collateral arbors are given in Table 2. Unlike the HFAs with their distinctive flame-shaped arbors, there is considerable variation in the morphology of rapidly adapting glabrous skin afferents. In the majority of arbors the collateral axon branches dorsal to the terminal arborization. In some cases, though, the collateral curves medially producing a cluster of terminal branches that lie adjacent to the medial edge of the dorsal horn and in others the terminal branches are issued at intervals as the collateral axon descends ventrally into the dorsal horn (Fig. 3). The distribution of the terminal boutons of these rapidly adapting afferents was more widespread than HFAs, the majority being located in laminae IV-V with some extending to lamina I11 and inner 11. Somatotopic location of terminals. The area occupied by the terminals of the eight glabrous skin rapidly adapting (RA) afferents, all of which had receptive fields on the toes, is shown in Figure 4. The terminals are distributed in the medial third of the dorsal horn from mid L4 to caudal L5. Laterally they border with the central terminal fields of the superficial peroneal, sural, and saphenous nerves. The total tibia1 terminal region is larger than this area as it includes paw pad afferents as well as toe afferents and extends more rostrally to the L3iL4 border and more caudally to the L4/L5 border (Swett and Woolf, ’85; Molander and Grant, ’86; Woolf and Fitzgerald, ’86).

Sproutingof saphenousHFAs following neonatal sciaticnerve section Somatotopic location and dimensions of sprouts. Eighteen individual saphenous nerve hair follicle afferents were filled with HRP in adult rats whose sciatic nerve had been cut at birth. Twelve of these were found to have grown bouton containing arbors into the area of cord normally exclusively occupied by the central complex and simple terminals of sciatic nerve afferents. These sprouts manifested as densely branching collateral arbors studded with boutons 1,000 km caudal to mid L4, the normal limit of saphenous bouton containing arbors. In three afferents the sprouts extended caudally to the mid L5 region (2,000 km caudal to mid L4, Fig. 2B). The extent of the caudal expansion of the saphenous terminal field can be seen in Figure 2B. Altogether a total of 41 of the 106 collaterals from these 12 afferents were found to lie within the sciatic terminal region and therefore represent sprouts. Both the area normally containing sciatic hairy afferents (H, in Fig. 2A) and that containing sciatic glabrous afferents (G,in Fig. TABLE 2. Dimensions of Glabrous Rapidly AdaptingMerent Terminals’ Control N Mean No. of collaterals Total Complex Simple Blind-ending Intercollateral distance, pm Total rostmcaudal collateral distance, mm Arbr dimensions, +m Mediolateral Complex Simple Dormventral Complex Simple Rostrocaudd Complex Simple ‘ x i SEM.

Tibial intact

8

7

10.1 ? 0.5 4.1 t- 0.5 2.8 t- 0.7 3.6 ? 0.5 289 ? 21

9.6 i 1.1 3.4 ? 0.5 2.9 i 0.4 3.3 i 1.0 391 i 29

2.9 rr 0.3

3.4 -i 0.4

99i6 66 ? 10

140 t 12 61 ? 8

259 2 21 157 ? 24

225 t- 21 107 = 18

210 ? 14 169 ? 13

216 ? 19 166 i 15

2A) become invaded by complex and simple collaterals of saphenous nerve terminals. An equal number of afferents (n = 6) sent sprouts into the glabrous zone as into the hairy zone of the sciatic territory. No change in the overall length of the terminal field of individual saphenous HFAs occurred even though they extended more caudally than normal. Table 1 shows that the mean distance between the most rostral and caudal collateral of saphenous HFAs in sciatic sectioned animals is not significantly different from controls. Six saphenous hair follicle afferents did not possess arbors in the sciatic territory. The complex terminal field of 5 of these afferents were restricted to the L3 segment and the 6th was largely located in L3 so that they were all some distance from the deafferented terminal zone. In Table 1 data from these six afferents have been separated from the 12 afferents that were more caudally positioned and did sprout into the sciatic terminal territory. The saphenous terminals also sprouted in the mediolatera1 plane. In this case, expansion of the overall terminal field has to be viewed in the context of a 25% reduction in the width of the dorsal horn because of the growth retardation of L4 that follows neonatal sciatic nerve section (Fitzgerald and Shortland, ’88). However, measurement of the absolute mediolateral widths of individual collateral arbors demonstrates that substantial real growth has occurred in this dimension. Table 1 shows that, overall, the mean widths of the complex arbors of saphenous HFAs in sciatic sectioned animals were 50% larger than in control animals. The expansion was greater (68%) in those afferents with collaterals in L4, many of which were in sciatic territory, but it was also considerable (33%)in afferent collaterals in L3 that were still confined to the saphenous terminal field. There was also a smaller (10%)dorsoventral increase in the size of complex arbors compared to controls. Morphology of the sprouts. The general appearance of the sprouted collateral arbors depended on whether they lay in the hairy or glabrous part of the sciatic denervated zone. Collaterals in that part of the L4 segment which normally contain hair follicle afferent terminals whether from the saphenous (in rostral L4) or sciatic (in caudal L4) nerves resembled the flame-shaped arbors of normal HFAs (Figs. 5, 6). These arbors were wider in the mediolateral plane than control HFAs and had branches outside what would be expected to be their normal terminal zone (Figs. 5, 6; Table 1).Out of the 41 collaterals that had sprouted into the sciatic territory, 12 (29%) fell into this group. The remaining 29 collaterals (71%) were located in the medial L4 and L5 dorsal horn, an area which normally receives glabrous skin afferents (Fig. 2A). These arbors, despite being the central terminals of physiologically identified HFAs, were not “flame-shaped” but had a widely varying morphology (Figs. 7, 8) that resembled glabrous skin afferent central terminals. Instead of forming a classical “U-shaped” dorsally directed collateral axon terminating in a tightly packed arbor densely studded with boutons, these collateral axons were ventrally directed, often with preterminal branches producing a widespread terminal arborization with a consequent decrease in the density of boutons. Some collaterals formed dorsal and ventral terminal branches pressed against the medial edge of the dorsal horn (Figs. 7, 8), a pattern that is normally characteristic of glabrous skin afferents in L4. Some afferents had collaterals with a morphology typical of HFAs and caudal collaterals with features of RA glabrous skin afferents (Fig. 7).

CENTRAL SPROUTING OF PRIMARY AFFERENTS

377

c Fig. 5. Camera lucida reconstruction of 6 adjacent terminal arborizations of a saphenous nerve HFA from an animal whose sciatic nerve was sectioned at birth. This afferent sprouted but remained within an area of cord normally devoted to hairy afferent inputs. Scale bar, 250 pm.

Sproutingoftibidnerve afferent followh@x?onatalsectionof the sciatic and saphenousnerves Location and dimensions of sprouted terminals. Seven tibial rapidly adapting glagrous skin fierents were

recorded from rats whose sural, common peroneal, and saphenous nerves had been cut at birth (denervating the hairy skin surrounding the glabrous skin of the hindpaw). All had receptive fieldson the toes. Four of the 7 had central terminals that underwent growth into new areas of the

M. FITZGERALD ET AL.

378

Fig. 6. A further example of all the collaterals from rostral to caudal (A-K) of a saphenous nerve HFA that has sprouted following neonatal nerve section into areas of cord normally containing hair afferents. Scale bar, 250 pm.

spinal cord. Twenty-five collaterals, out of a total of 67 (37%), sprouted in the mediolateral dimension, 12 (48%) into areas normally devoted only to hairy skin terminals. Quantitative details of the growth are provided in Table 2. Unlike the HFAs, glabrous afferents did become elongated with the distance from the most rostral to the most caudal collateral increasing by 16%. Since the number of collaterals stayed the same, the result was an increase in intercollateral spacing. The most striking feature of these afferents, though, was their mediolateral growth producing a lateral extension of the tibial terminal field. Even taking into account the 20%shrinkage of the L4 dorsal horn that results from sural nerve section, the lateral expansion of the tibial field is evident in Figure 4B.Table 2 shows that individual terminal arbors in tibial intact rats were 40% larger in the mediolateral dimension than control afferents.

Morphology of the sprouts. As in the case of sprouted HFA terminals, the morphology of the terminals of sprouted glabrous RA afferents depended on their central location. Those that remained in the tibial nerve terminal region retained the appearance of normal, if wider, RA glabrous skin afferents (Fig. 91, while those glabrous skin afferents that sprouted into “hairy” regions of cord took on the characteristic and unmistakable appearance of an HFA (Fig. 10).

Peripheral receptivefields The receptive fields of the saphenous HFAs in neonatally sciatic sectioned animals were normally sized and all lay within the saphenous skin territory. There was no physiological evidence of any peripheral sprouting into the dener-

CENTRAL SPROUTING OF PRIMARY AFFERENTS

3 79

a X K

Figure 6 continued

vated sciatic skin region. This was supported by silver staining of hair follicle terminals in the skin which also failed to reveal any peripheral sprouting of saphenous skin terminals into denervated sciatic skin (in preparation). Glabrous skin RA f i e r e n t s in the tibial intact rats also had normal receptive fields restricted to the normal tibial skin territory. In other words, as illustrated in Figures 2 and 4, the presence of central sprouting following neonatal deafferentation is not related to any change in the location of peripheral receptive fields.

DISCUSSION Central collateral sprouting of intact saphenous nerve terminals in the rat spinal cord following neonatal sciatic

nerve section has been shown previously by bulk labelling of the whole saphenous nerve with HRP (Fitzgerald, '85). A very similar response of trigeminal afferents has also been observed following prenatal infraorbital nerve section (Rhoades et al., '89). These observations, while establishing the capacity of afferent terminals to expand into denervated areas, provide no detail of the pattern and organization of the response at the single afferent level or of the morphology of the new growth. The present study provides this information and goes further in providing evidence of an influence of central targets on afferent terminal morphology. Single saphenous HFAs can grow new bouton containing terminals up to 2,000 pm caudal to their normal terminal field, into the deafferented sciatic terminal region

380

M. FITZGERALD ET AL.

Fig. 7. Camera lucida reconstruction of 6 adjacent terminal arborizations from rostra1 to caudal (A-F) of a saphenous nerve HFA from an animal whose sciatic nerve was sectioned at birth. The first two collaterals are still within the saphenous nerve territory and have

flame-shaped arbors but collaterals C-F have sprouted into an area of cord normally devoted to glabrous skin aEerents and have taken on the appearance of glabrous skin mechanoreceptors. Scale bar, 250 km.

and single tibial glabrous RA afferents can grow into the saphenous, superficial peroneal and sural terminal regions. There are a number of different ways in which a n axon may sprout. One involves growth or expansion of existing terminals to cover wider regions (Garraghty et al., '88;

Renehan et al.,'89). This was observed here in the expansion of the flame-shaped arbors of some HFAs within the normal saphenous region following sciatic nerve section. This was also the response of most tibial afferents which expanded mediolaterally into surrounding deafferented re-

CENTRAL SPROUTING OF PRIMARY AFFERENTS

381

Fig. 8. A further example of a saphenous nerve HFA that has sprouted following neonatal sciatic nerve section. In this case all the illustrated collaterals (A-F) are in the area normally devoted to glabrous skin afferents and have acquired the appropriate morphology. Scale bar, 250 pm.

gions of cord. Another form of growth might be the elaboration of terminals on blind-ending collaterals (Wilson and Snow, ’88) and this is likely to account for some of the rostrocaudal expansion of terminal fields observed here. A third form of sprouting would involve the formation of new terminals in areas where the afferent nerve normally extends. This form of growth was also observed here in the appearance of new saphenous HFA collaterals beyond even the normal caudal extent of the blind-ending terminals. This sprouting represents new growth and not a failure of withdrawal of excess terminals formed earlier in development, since primary afferents grow into specific areas of

cord (Smith, ’83; Fitzgerald, ’87) and do not display the “exuberant” growth seen in some parts of the CNS. Neonatal sciatic nerve section results in the death of the majority of sciatic dorsal root ganglion cells (Yip et al., ’84; Bondok and Sansone, ’84) and therefore results in a massive deafferentation of the medial L4 and L5 dorsal horn. In the adult rat, where nerve section does not cause such an extensive dorsal root ganglion loss (Devor et al., ’85)sprouting does not occur following simple sciatic nerve section (Stelzner and Devor, ’84; Fitzgerald, ’85;Molander et al., ’88).If adult dorsal root ganglion cells are destroyed, however, by, for example, pronase injection into the sciatic

Figure 9

CENTRAL SPROUTING OF PRIMARY AFFERENTS

383

A

Fig. 10. A further example of a glabrous skin toe rapidly adapting mechanoceptor that has sprouted following neonatal “tibial intact” deafferentation. In this case, the illustrated collaterals are all in areas of cord normally devoted to “hairy” skin afferents and have taken on a flame-shaped morphology characteristic of HFAs. Scale bar, 250 p.m.

nerve, then the saphenous nerve terminal field does expand denervated skin has been described in adult rat (Devor et (Lamotte et al., ’89). The trigger for collateral sprouting al., ’79, Jackson and Diamond, ’84;Kinmann, ’87) and appears to be therefore the creation of a considerable transiently in the newborn (Kinmann and Aldskogius, ’86) number of vacant postsynaptic sites. It appears to be but no evidence of growth of the large myelinated sapheunrelated to peripheral sprouting. Collateral sprouting into nous or tibial terminals into neighbouring denervated skin was observed in the present experiments. The extent of sprouting we observed may be related to the fact that the intact saphenous fierents were still in a growing mode at Fig. 9. Camera lucida reconstructions of 6 adjacent terminal ar- the time of the deafferentation, and therefore would have borizations from rostral to caudal (A-F) of a glabrous skin toe rapidly all the metabolic machinery for the extension of neurites adapting mechanoceptor in an animal whose tibial nerve only was left and the formation ofsynapses. Peripheral collateral sproutintact at birth. This afferent has sprouted in the mediolateral plane is by anti-nerve growth factor hut stayed largely within the normal glabrous skin region of the cord. ing Of One branch, collateral (D), has grown laterally into hairy skin region (NGF) (Diamond et al.7 1987); and it is Possible, Since the central processes of dorsal root ganglion (DRG) cells are and acquired a flame shape. Scale bar, 250 pn.

384

known to have NGF receptors (Richardson et al., 19861, that NGF also plays a role in the central sprouting reported here. If neonatal rats are treated with capsaicin, a C-fibre neurotoxin, this results in a selective death of the majority of small diameter afferents (Jansco et al., '77) with a consequent denervation of the superficial dorsal horn, their site of central termination (Sugiura et al., '86). We have previously shown that in this situation, the arbors of HFAs grow into the substantia gelatinosa and their total dorsoventral length increases (Shortland et al., '90).This represents a different form of sprouting to that described here in that the somatotopic organization of the afferents is maintained but their laminar distribution is altered. The caudal growth of saphenous terminals into the sciatic region was not associated with an increase in the length of the afferents. While this may be an artefact of dye filling, it may mean that the new collateral developed at the expense of more rostral collaterals. At the time of the sciatic deafferentation, the day of birth, saphenous HFAs have grown into the cord and formed flame-shapedarbors (Smith, '83; Beal et al., '88; Fitzgerald et al., submitted) but considerable growth is still in progress and a potential for a change in existing collaterals is possible. Support for the idea of withdrawal of some rostral arbors comes from the fact that those arbors that remain have enlarged terminal arborizations despite the fact that they are some distance from the deafferented sciatic zone. There may be an optimum size of terminal field that can be supported by a single primary afferent that is not only achieved in normal development but maintained during plastic reorganization of the spinal cord. The sprouted collateral terminals did not necessarily have the characteristic morphology expected of HFAs or RA glabrous afferents. Instead we have found that if the collateral terminated in areas of cord normally occupied by glabrous skin afferents they resembled glabrous skin afferents, but if they sprouted into areas normally occupied by hairy skin d e r e n t s then they resembled hair follicle afferents. This was independent of their receptive field location, physiological receptor type, and terminal morphology in the normal target region. These results show that the terminal growth patterns of new collateral sprouts are governed by central rather than peripheral targets. The idea of target and environmental control over neuronal phenotype is not a new one. It has been demonstrated during development in neural crest cells (Le Douarin, '71) and sympathetic neurons (Landis and Keefe, '83; Potter et al., '86) and in the adult during muscle (Bennet et al., '73) and Pacinian corpuscle graft (Zelena and Jirmanova, '88) reinnervation. In the central nervous system retinal terminals have been shown to develop a lemniscal terminal-like morphology and ultrastructure when induced to take up permanent residence in the ventrobasal nucleus of the thalamus instead of the lateral geniculate nucleus (Campbell and Frost, '87). Our results relate to mechanisms of plasticity, but it is possible that similar mechanisms operate in normal development and that, for instance, growing hair follicle afferents are induced to form "flame" arbors by some aspect of their target in the dorsal horn. Consistent with this is the finding in the cat that hair follicle afferent terminals in the cmeate nucleus have a different morphology from those in the dorsal horn (Fyffe et al., '86) as do the terminals of muscle afferents in Clarke's nucleus compared to those in the ventral horn (Hongo et al., '87). The development of

M. FITZGERALD ET AL. terminal morphology would therefore differ from earlier events in primary afferent growth where collaterals first enter the spinal cord and find their correct termination zone in a process that is influenced, at least in the frog, by peripheral cues (Smith and Frank, '88). It is likely that a combination of peripheral and central influences may act to ensure both the correct general arrangement and the specific morphology of the central terminals of primary sensory neurones.

ACKNOWLEDGMENTS We thank the MRC and the Wellcome Trust for financial support and P. Ainsworth and J. Middleton for technical assistance.

Beal, J.A., D.S. Knight, and K.N. Nandi (1988) Structure and development of central arborizations and hair follicle primary afferent fibres. Anat. Embryol. (Berl.) 178t271-279. Bennet, M.R., A.G. Pettigrew, and R.S. Taylor (1973) The formation of synapses in reinnervated and cross-reinnervated adult avian muscle. J. Physiol. (Lond.) 23Ot331-357. Bondok, A.A., and F.M. Sansone (1984) Retrograde and transganglionic degeneration of sensory neurons after a peripheral nerve lesion at birth. Exp. Neurol. 86:322-330. Brown, A.G., P.K. Rose, and P.J. Snow (1977) Morphology of hair follicle afferent fibre collaterals in the cat spinal cord. J. Physiol. (Lond.) 272:779-797. Campbell, G., and D.O. Frost (1987) Target-controlled differentiation of axon terminals and synaptic organization. P.N.A.S. 845929-6933. Cotman, C.W., M. Nieto-Sampedro, and E.W. Harris (1981) Synapse replacement in the central nervous system of adult vertebrates. Physiol. Rev. 61:684-784. Devor, M., D. Schonfeld, Z. Seltzer, and P.D. Wall (1979) Two modes of cutaneous reinnervation following peripheral nerve injury. Nature 185; 211-220. Devor, M., R. Govrin-Lippmann, I. Frank, and P. Raber (1985) Proliferation or primary sensory neurons in adult rat DRG's and the kinetics of retrograde cell loss after sciatic nerve section. Somatosensory Res. 3:139-167. Diamond, J., M. Coughlin, L. Macintyre, L.M. Holmes, and B. Visheau 11987) Evidence that endogenous p-nerve growth factor is responsible for the collateral sprouting, but not the regeneration of nociceptive axons in adult rats. P.N.A.S. 84:659&6600. Fitzgerald, M. (1985) The sprouting of saphenous nerve terminals in the spinal cord following early postnatal sciatic nerve section in the rat. J. Comp. Neurol. 240:407-413. Fitzgerald, M. (1987) The prenatal growth of fine diameter afferents into the rat spinal cord-a transganglionic study. J. Comp. Neurol. 261:98-104. Fitzgerald, M., M. Reynolds, and L.I. Benowitz. The development of the rat lumbar spinal cord: a GAP-43 immunocytochemical study (submitted). Fitzgerald, M., and P. Shortland (1988) The effect of neonatal peripheral nerve section on the soma dendritic growth of sensory projection cells in the rat spinal cord. Dev. Brain Res. 42:129-136. Fitzgerald, M., and C.J. Woolf (1987) Reorganization of HFA terminals in the rat spinal cord following peripheral nerve section. Neuroscience 22:2380P. Fyffe, R.W., S.S. Cheema, and A. Rustioni (1986) Intracellular stainingstudy of the feline cuneate nucleus. I. Terminal patterns of primary afferent fibres. J. Neurophysiol. 56:1268-1283. Garraghty, P.E., C.J. Shatz, D.W. Sretavan, and M. Sur (1988) Axon arbors of X and Y retinal ganglion cells are differentially affected by prenatal disruption of binocular inputs. P.N.A.S. 85;1-7365. Guillery, R.W. (1972) Experiments to determine whether retinogeniculate axons can form translaminar collateral sprouts in the dorsal lateral geniculate nucleus of the cat. J. Comp. Nenrol. 146;407-420. Hanker, J.S., D.E. Yates, C.B. Metz, and A. Rustioni (1977) A new specific and non-carcinogenic reagent for the demonstration of HRP. d. Histochem. 9:789-792. Hongo, T., N. Kudo, S. Sasaki, M. Yamashita, K. Yoshia, N. Ishizuka, and H.

CENTRAL SPROUTING OF PRIMARY AFFERENTS Mannen (1987) Trajectory of group Ia and Ib fibres from hindlimb muscles at the L3 and L4 segments of the spinal cord of the cat. J. Comp. Neurol. 262159-194. Jackson, P.C., and J. Diamond (1984) Temporal and spatial constraints on the collateral sprouting of low-threshold mechanosensory nerves in the skin of rats. J. Comp. Neurol. 226:336-345. Jansco, G., E. Kivaly, and A. Jansco-Gabor (1977) Pharmacologically induced selective degeneration of chemosensitive primary sensory neurons. Nature 270:741-743. Kalil, R.E., and G.E. Schneider (1975) Abnormal synaptic connections of the optic tract in the thalamus after midbrain lesions in newborn hamster. Brain Res. 100:690-698. Kinmann, E. (1987) Collateral sprouting of sensory axons in the hairy skin of the trunk: a morphological study in adult rats. Brain Res. 414:385-389. Kinnman, E., and H. Aldskogius (1986) Collateral sprouting of sensory axons in the glabrous skin of the hindpaw after chronic sciatic nerve lesion in adult and neonatal rats: a morphological study. Brain Res. 377:73-82. Lamotte, C.C., S.E. Kapadia, and C.M. Kocol(1989) Deafferentation induced expansion of saphenous nerve terminal field labelling in the adult rat dorsal horn following pronase injection of the sciatic nerve. J. Comp. Neurol. 288:311-325. Landis, S.C., and D. Keefe (1983) Evidence for neurotransmitter plasticity ‘in vivo’: Developmental changes in the properties of cholinergic sympathetic neurons. Dev. Biol. 98:349-372. Le Douarin, N.M. (1971) The ontogeny of the neural crest in avian embryo chimaeras. Nature 186:663-669. Meyers, D.E.R., and P.J. Snow (1984) Somatotopically inappropriate projections of single HFA’s to cat spinal cord. J. Physiol. (Lond.) 347:59-73. Molander, C.M., and G. Grant (1986) Lamina distribution and somatotopic organization of primary afferent fibres from hindlimb nerves in the dorsal horn: study by transganglionic transport of WGA-HRP conjugate. Neuroscience 19:29 7-3 12. Molander, C., E. Kinnman, and H. Aldskogius (1988) Expansion of spinal cord primary sensory afferent projection following combined sciatic nerve crush. J. Comp. Neurol. 276:436-441. Potter, D.D., S.G. Matsumoto, S.C. Landis, D.W.Y. Sah, and E.J. Furshpan (1986) Transmitter status in cultured sympathetic principal neurons: plasticity, graded expression and diversity. Prog. Brain Res. 68:103-127. Raisman, G., and P.M. Field (1973) A quantitative investigation of the development of collateral reinnervation after partial deafferentation of the septa1 nuclei. Brain Res. 33r471-476. Renehan, W.E., B.G. Klein, N.L. Chiaia, M.F. Jacquin, and R.W. Rhoades (1989) Physiological and anatomical consequences of infraorbital nerve transection in the trigeminal spinal tract of the adult rat. J. Neurosci. 9:548-557. Rhoades, R.W., N.L. Chiaia, G.J. McDonald, and M.F. Jacquin (1989) Effect of fetal ION transection upon trigeminal primary afferent projections in the rat. J. Comp. Neurol. 287:82-97. Richardson, P.M., V.M.K. Verge Issa, and R.J. Riopelle (1986) Distribution of neuronal receptors for nerve growth factor in the rat. J. Neurosci. 6.2312-2321.

385 Rodin, B.E., and L. Kruger (1984) Absence of intraspinal sprouting in dorsal root axon6 caudal to a partial spinal hemisection-HRP transport study. Somatosens. Res. 2171-192. Rodin, B.E., S.L. Sampogna, S.L., and L. Kruger (1983) An examination of intraspinal sprouting in dorsal root axons with the tracer HRP. J. Comp. Neurol. 215:187-198. Rustioni, A,, and I. Molenaar (1975) Dorsal column nuclei afferents in the lateral funiculus of the cat: distribution pattern and absence of sprouting after chronic deafferentation. Exp. Brain Res. 23:l-13. Shortland, P., C. Molander, C.J. Woolf, and M. Fitzgerald (1990) Neonatal capsaicin treatment induces invasion of the substantia gelatinosa by the terminal arborizations of hair follicle afferents in the rat dorsal horn. J. Comp. Neurol. 296t23-31. Shortland, P., C.J. Woolf, and M. Fitzgerald (1989) Morpholoa and somatotopic organization of central terminals of hindlimb hair follicle afferents in the rat lumbar spinal cord. J. Comp. Neurol. 289:416433. Smith, C.L. (1983) The development and postnatal organization of primary afferent projections to the rat thoracic spinal cord. J. Comp. Neurol. 220:29-43. Smith, C.L., and E. Frank (1988) Specificity of sensory projections to the spinal cord during development in bullfrogs. J. Comp. Neurol. 269:96108. Stanfield, B., and W.M. Cowan (1976) Evidence for a change in the retino-hypothalamic projection in the rat following early removal of one eye. Brain Res. 104:129-136. Stelzner, Z., and M. Devor (1984) Effect of nerve section on the distribution of neighbouring nerves. Brain Res. 3063-37. Sugiura, Y., E.R. Perl, and C.L. Lee (1986) Central projections of identified unmyelinated (C) afferent fibres innervating mammalian skin. Science 234:358-361. Swett, J.E., and C.J. Woolf (1985) Somatotopic organization of primary afferent terminals in the superficial laminae of the dorsal horn of the rat spinal cord. J. Comp. Neurol. 231:66-77. Tsukahara, N. (1981)Synaptic plasticity in the mammalian central nervous system. Annu. Rev. Neurosci. 4:351-379. White, E.L., and F.D. Nolan (1974) Absence of re-innervation in the chinchilla medial superior olive. Anat. Rec. I78:486,487. Wilson, P. and P.J. Snow (1988) Alterations of dorsal horn somatatopy after neonatal peripheral nerve section may result from collateral sprouting by intact primary afferent fibres. Proc. Aus. Physiol. & Pharmacol. SOC. 19.181 p. Woolf, C.J. (1987) Central terminations of cutaneous rnechanoreceptive afferents in the rat lumbar cord. J. Comp. Neurol. 261:105-119. Woolf, C.J., and M. Fitzgerald (1986) The somatotopic organization of cutaneous afferent terminals and dorsal horn neuronal receptive fields in the superficial and deep laminae of the rat lumbar spinal cord. J. Comp. Neurol. 251:5 17-53 1. Woolf, C.J., and A.E. King (1987) Physiology and morphology of multireceptive neurons with C-afferent inputs in the deep dorsal horn of the rat lumbar spinal cord. J. Neurophysiol. 58;460479. Yip, H.K., K.M. Rich, P.A. Lampe, and E.M. Johnson (1984) The effects of nerve growth factor and its antiserum on the postnatal development and survival after injury of sensory neurons in rat dorsal root ganglia. J. Neurosci. 4:2986-2992. Zelena, J.,and I. Jirmanova (1988) Grafts of pacinian corpuscles reinnervated by dorsal root axons. Brain Res. 438:165-174.

Collateral sprouting of the central terminals of cutaneous primary afferent neurons in the rat spinal cord: pattern, morphology, and influence of targets.

The capacity of the central terminals of primary afferents to sprout into denervated areas of neonatal spinal cord and the morphology of any novel ter...
1MB Sizes 0 Downloads 0 Views