DEVELOPMENTAL

BIOLOGY

147,39%402 (1991)

Axonal Transport and Release of Transferrin in Nerves of Regenerating Amphibian Limbs WILLIAM Medical

R. KIFFMEYER,ERIK

Sciences

Program,

V. TOMUSK,ANDANTHONY

Indiana University Accepted

School July

of Medicine,

Bloomington,

L. MESCHER~ Indiana

47.405

3, 1991

Transferrin, a plasma protein required for proliferation of normal and malignant cells, is abundant in peripheral nerves of birds and mammals and becomes more concentrated in this tissue during nerve regeneration. We are testing the hypothesis that this factor is involved in the growth-promoting effect of nerves during the early, avascular phase of amphibian limb regeneration. A sensitive enzyme-linked immunosorbent assay for axolotl transferrin was developed and used to determine whether this protein meets certain criteria expected of the trophic factor(s) from nerves. During limb regeneration adult sciatic nerves greatly increased their content of transferrin, which immunohistochemistry revealed was distributed in both axons and Schwann cells. Using the double ligature method with sciatic nerves in vivo, it was determined that transferrin is carried by fast anterograde axonal transport at all stages of limb regeneration. An approach based on multicompartment organ culture demonstrated that fast-transported transferrin was secreted in physiologically significant amounts at distal ends of regenerating axons. Finally, the concentration of transferrin in the distal region of larval axolotl limb stumps was found to decrease directly and rapidly in response to axotomy. Since transferrin is important for both axonal regeneration and cell cycling, the present data have significance for various o 19% Academic PWS, I~C. aspects of nerve’s trophic activity during limb regeneration.

bus and Vethamany-Globus, 1985). Singer (1978) has argued that these factors are present in all neurons as well as other cells and are important for growth and maintenance of axonal processes. This view is supported by recent work demonstrating increased growth-promoting activity as nerves themselves regenerate (Boilly and Albert, 1988). Numerous studies on effects of denervation indicate that nerves maintain the metabolic and proliferative rates of blastema cells, with little or no qualitative effect on cellular activity or gene expression (reviewed by Singer, 1978; Wallace, 1981). Muneoka et al. (1989) have argued that the neural influence involves a permissive factor required by blastema cells to sustain proliferative activity, rather than a regulatory factor which signals cells to begin the cell cycle. Similar views have also been expressed by others (Globus, 1978; Singer, 1978; Goldhammer and Tassava, 1987). One such nonregulatory growth factor is transferrin, the plasma protein for delivery of iron to cells of vertebrates. Transferrin is required for proliferation of nearly all normal and transformed cells in defined, serum-free medium (Barnes and Sato, 1980) and is important primarily for progression through the cell cycle (Pardee et al., 1981). Rapidly dividing cells have high levels of cell-surface transferrin receptors that are expressed in response to mitogenic stimuli (May and Cuatrecasas, 1985; Davis and Czech, 1986; Roberts et ab, 1989). Blocking uptake of transferrin/iron with monoclonal antibodies against the receptor rapidly inhibits cell cycling (Seligman, 1983). Transferrin appears to be re-

INTRODUCTION

Work with polypeptide growth factors has contributed greatly to our understanding of how injury elicits tissue repair and wound healing (Clark and Henson, 1988). The cellular events that follow amputation of urodele amphibian limbs have long been of special interest to developmental biologists because here the injuryinduced proliferation produces a population of mesenchymal cells, the blastema, out of which the missing structures of the limb develop. Blastema formation and growth require other local factors in addition to those released as a direct result of the tissue injury. The epithelium which migrates from the epidermis to cover the amputation wound appears to function like the apical ectoderm of embryonic limb buds in promoting distal outgrowth of the underlying mesenchymal cells (Tassava and Mescher, 1975; Wallace, 1981). Unlike limb buds, blastema growth also requires the presence of axons from nerves in the limb stump (Singer, 1952,1978; Brockes, 1987). The trophic effect of nerves during limb regeneration is understood in broad terms, but the nature of the substance(s) involved remains elusive. Considerable evidence supports the concept that axons transport and release diffusible polypeptide factors needed for cell division in the blastema (reviewed by Wallace, 1981; Glo’ To whom correspondence should be addressed. 0012-1606/91$3.00 Copyright All rights

0 1991 by Academic Press, Inc. of reproduction in any form reserved.

392

KIFFMEYER,

TOMUSK,

AND MESCHER

quired to provide iron needed as a cofactor for many enzymes, including the rate-limiting enzyme for DNA synthesis (Seligman, 1983; Bowman et al., 1988; Reichard, 1988; de Jong et al, 1990). We became interested in nerves as a possible source of transferrin for blastema cells when it was found that peripheral nerves are particularly rich in this factor (Beach et ab, 1983; Markelonis et al., 1982; Meek and Adamson, 1985). Initial studies showed that heterologous transferrin stimulates mesenchymal cell proliferation in cultured blastemas from newts (Mescher and Munaim, 1984) and axolotls (Albert and Boilly, 1988). In both investigations the transferrin dose response was biphasic, which has also been reported with various mammalian cells (Barnes and Sato, 1980; Perez-Infante and Mather, 1982; Seligman, 1983). A high transferrin concentration with reduced activity becomes fully effective in the presence of additional iron salts, as observed both with mammalian cells (Perez-Infante and Mather, 1982) and with amphibian blastema cells (Mescher and Munaim, 1984; Albert and Boilly, 1988). These features of the transferrin dose response are of interest because extracts of neural tissue have also been found repeatedly to show biphasic responses when tested for growth-promoting activity on blastema cells in vitro (Choo et al, 1978; Mescher and Loh, 1981; Carlone and Rathbone, 1985; Munaim and Mescher, 1986; Boilly and Albert, 1988). Although the diminished activity of such extracts at high doses may have various explanations, a high concentration of neural extract stimulates blastema growth if additional iron is provided (Munaim and Mescher, 1986). More important, the growth-promoting activity of the neural extract is removed by dialysis with an iron-chelating agent and is completely restored by readdition of iron (Munaim and Mescher, 1986). Although these results strongly suggest that the presence of transferrin is required for the stimulatory effect of nervous tissue extracts on cultured blastema cells, their relevance to the influence of nerves during regeneration in vivo remains to be shown. The objective of the present study was to answer certain questions relevant to the hypothesis that transferrin is involved in the trophic effect of nerves on blastema growth in vivo (Mescher and Munaim, 1988). Specifically, experiments were designed to determine whether transferrin levels increase in nerves during regeneration, whether this protein is distally transported in axons and released, and whether it is lost from the blastema following denervation. MATERIALS

AND

METHODS

SurgicaZ Procedures

Adult and larval (5-6 cm) axolotls (Ambystoma meziobtained from the Indiana University Axolotl

canum)

Tran.$errin

in Regenerating

Nerves

393

Colony were used in all experiments. Animals were maintained in 25% Holtfreter’s solution at 20°C and fed brine shrimp or pellets of trout food every other day. All surgical procedures were performed after anesthesia by immersion in 0.02% benzocaine (ethyl paminobenzoate; Sigma) in 25% Holtfreter’s solution. Forelimbs of larval axolotls were amputated bilaterally distal to the elbow. One limb of each animal was denervated with fine forceps by transection of brachial nerves 3,4, and 5 near the brachial plexus at the time of amputation, 3 or 5 days postamputation. The contralatera1 limb served as the innervated control. Six days postamputation the distal 0.5 mm of forelimb tissue was removed from each limb stump (excluding limbs with visible blood clots below the apical epithelium), homogenized in a Wheaton glass microtissue grinder in 1.0 ml of buffer containing 10 milf KCl, 1.0 mM MgCl,, 0.1 mM phenylmethylsulfonyl fluoride (pH 7.5), and assayed for transferrin as described below. Forelimbs of another group of larvae were X irradiated unilaterally with 2000 rad using a Torrex 150 instrument and amputated bilaterally distal to the elbows 1 day later. All irradiated limbs failed to form blastemas while all the shielded limbs regenerated normally. Six days postamputation the distal 0.5 mm of each limb stump was collected, homogenized as described, and assayed for transferrin. Hindlimbs of adult axolotls were unilaterally amputated distal to the knee, the contralateral limb serving as the unamputated control. Limbs were permitted to regenerate to either the early bud, early digit, or late digit stage at which times the sciatic nerves of both limbs were ligated with 4-O silk at two places 9 mm apart (Tetzlaff et ak, 1989). Three hours later each nerve, totalling 18-20 mm in length with at least 6 mm above the proximal ligature and 3 mm beyond the distal ligature, was removed and the animal euthanized. After removing external connective tissue and large blood vessels, each nerve was cut into 3-mm segments which were homogenized as described and assayed for transferrin. Nerves in which blood had collected visibly within the perineurium were not used. Rates of anterograde and retrograde transport and the mobile fraction of transferrin were determined from the immunoassay data using equations described by Rasool et al. (1981) and Tetzlaff et al. (1989). Colchicine (Sigma) was embedded in the slow-release polymer Elvax 40 (a gift of R. Langer) and 5-10 1.18was applied locally at the sacral plexus of regenerating limbs of adult axolotls using the procedure of Scadding (1988). Three days after implanting colchicine the sciatic nerves were doubly ligated and axonal transport determined as described above.

DEVELOPMENTALBIOLOGY VOLUME147,1991

394 Immunoassay

and Immunolocalixation

of Transferrin

Adult axolotls were exsanguinated by cardiac puncture and serum was collected after clotting and centrifugation. Purification of transferrin followed the procedure of Werner et al. (1983). Serum was dialyzed against Tris buffer (0.02 M Tris-HCl, 0.028 M NaCl, 0.02% sodium azide, pH 8.0) and passed over an affinity column of Affi-gel blue (Bio-Rad), followed by gel filtration over a column of Sephadex G-100 Superfine. Purity of the protein was determined by SDS-polyacrylamide gel electrophoresis. Rabbit antiserum against axolotl transferrin was prepared exactly as described by Meek and Adamson (1985). Female BALB/c mice were each immunized with 25 pg axolotl transferrin in 0.4 ml Freund’s complete adjuvant intraperitoneally and by subcutaneous injection at axillae. Two weeks later this procedure was repeated with antigen in Freund’s incomplete adjuvant. Titers of antisera from the rabbit and mice were determined by an enzyme-linked immunosorbent assay (ELISA) on microtiter plates. The immunoassay to determine transferrin levels in axolotl tissues and nerve organ culture medium was a noncompetitive sandwich-type ELISA similar to that used by Meek and Adamson (1985). Microtiter plates were coated with rabbit anti-transferrin serum (diluted 1:1500), followed by addition of sample homogenates, conditioned medium, or transferrin standards. Mouse anti-transferrin serum (diluted l:lOOO-2000) was added to the wells, followed by the secondary goat anti-mouse IgG conjugated to alkaline phosphatase (Sigma, diluted 1:500) and the p-nitrophenyl phosphate (Sigma) substrate. The plate was washed extensively with Tris buffer (20 mM Tris, 140 mM NaCI, 2.7 mM KCI, 0.05% Tween, pH 7.4) between steps. The optical densities were read at 405 and 540 nm on a Bio-Tek plate reader and transferrin concentrations determined from the standard curve. The ELISA was sensitive to 1 ng transferrin per milliliter. Sciatic nerves for immunolocalization of transferrin were removed from unamputated adult axolotls 2 hr after perfusion with 4% formaldehyde, 0.1% glutaraldehyde. Fascicles in 3-mm nerve segments were teased apart with fine forceps, after which the tissue was incubated for 5 hr in blocking solution (5% DMSO and goat serum diluted 1:30 in phosphate-buffered saline [PBS]). Following three 30-min washes in PBS, nerve segments were incubated overnight in a 1:500 dilution of either rabbit antiserum to axolotl transferrin or preimmune serum. After washing as before tissues were incubated 1 hr in FITC-conjugated goat anti-rabbit IgG (Sigma) diluted 1:500. Other nerve segments were treated similarly, but with a monoclonal antibody to neurofilament

protein (Boeringer-Mannheim) and rhodamine-conjugated goat anti-mouse IgG (Sigma). Nerves were photographed with a Nikon Fluophot microscope. In Vitro Analysis Transfer+

of Axonal

Transport

and Release of

Axonal regeneration was stimulated bilaterally by crush injuries to adult axolotl sciatic nerves at the knees. Two to four weeks later each nerve was dissected from the limb, including ganglia and the site of the conditioning lesion. Each nerve was cleaned of extraneous connective tissue and vessels, and placed in a multicompartment culture slide (Labtek) inside a lo-cm culture dish. All chambers were filled with serum-free Liebovitz L-15 (GIBCO), diluted to 80% and containing GIBCO antibiotic-antimycotic. For each nerve, ganglia were placed in one chamber and the distal portion containing the regenerating growth cones was placed in a second chamber. Tissue draped across the chamber wall was covered with petroleum jelly to prevent desiccation. The intervening region of nerve, approximately ‘7 mm in length, was bathed in medium of the culture dish between the two chambers, thereby preventing passive diffusion of material between chambers along the nerve. After different periods of culture (typically 24 hr) in a humidified chamber at 22”C, medium from the proximal and distal chambers was assayed for transferrin as described above. As a control for axonal transport, colchitine (0.5 mM) was added to the medium of the proximal chamber in some nerve cultures. RESULTS

Transferrin Concentration Regeneration

Increases in Nerve during

Because nerves regenerating after an earlier injury are reported to have greater growth-stimulating activity for blastemas than nerves not previously stimulated (Boilly and Albert, 1988; Maier et ab, 1984), we examined the effect of regeneration on the transferrin content in peripheral nerves. A sensitive, noncompetitive ELISA for axolotl transferrin showed that concentrations of this protein in adult sciatic nerves did increase substantially as limbs regenerated following unilateral amputation through the zeugopodium (Table 1). The mean transferrin content in sciatic nerves of unamputated animals was 9.1 rig/mm, but by the early bud stage of limb regeneration (approximately 4 weeks postamputation) the neural transferrin concentration had increased threefold. Accumulation of this protein in sciatic nerve continued, reaching much higher levels during the period of histogenesis and digit formation approximately 10 weeks postamputation. Interestingly,

KIFFMEYER,

TOMLJSK, AND MESCZHER

TABLE 1 TRANSFERRIN CONCENTRATIONS GENERATING AND CONTRALATERAL

IN SCIATIC NERVES FROM RECONTROL LIMBS OF ADULT Axo-

Lm?.s Stage of regeneration’

Regenerating (rig/mm)*

Early bud Palette Digital outgrowth

28.1 -+ 4.0 (4) 62.7 f 11.4 (4) 213.7 + 49.4 (3)

Control (rig/mm)” 14.3 + 2.4 (4) 54.4 + 13.9 (4) 159.3 + 19.1 (3)

a Staging system of Tank et al. (1976). *Mean + SE (n).

intact nerves in contralateral control limbs also showed increased transferrin concentrations, although in no case was the accumulation as high as in the regenerating nerves (Table 1). To reveal the cellular localization of this protein in peripheral nerves, preliminary indirect immunocytochemistry with peroxidase-conjugated antibodies was performed on unfixed frozen sections of adult sciatic nerve. Transferrin was clearly found in the perineurium by this method, but staining in axons and Schwann sheaths was less certain (unpublished observations). To localize transferrin more carefully within the myelinated fibers of sciatic nerve, tissue was fixed in situ by perfusion and indirect immunofluorescent light microscopy was performed on teased wholemount nerve preparations. As shown in Fig. 1, cells of the Schwann sheath and axons from which the myelin sheath had been removed were both stained with antiserum against the plasma protein. As others have found with sciatic nerves from chicken (Oh et aZ., 1981) and rat (Lin et al., 1990), staining in the Schwann cells appeared heavier in the outer margin of the myelin sheath and at the nodes of Ranvier, suggesting a cytoplasmic location. Identification of bare axons was confirmed by staining with a monoclonal antibody against neurofilament protein and all such axons were found to contain transferrin.

Transfer% Transport

Undergoes Fast Anterograde

Axonal

To test the possibility that transferrin is transported in axons of sciatic nerves in regenerating limbs, the ELISA was combined with a standard double-ligature technique for measuring axonal transport. Transferrin was found to accumulate in nerve segments proximal to the collection ligatures, in 3 hr reaching levels 2.5-times greater than the mean concentration in the segments between the ligatures (Fig. 2), a result consistent with transport of the protein in the anterograde direction.

Transferrin

in Regenerating Nerves

395

Accumulation distal to the ligatures, indicative of retrograde transport, was not observed. Three observations indicate that the accumulation of this plasma protein at the ligature is not the result of vascular constriction, but is due rather to microtubuledenendent axonal transnort. Transferrin collected onls on the proximal side of the ligatures and not symetritally, as is expected when accumulation is due to capillary constriction, edema, or other local artifact (Bisby, 1982). Perfusion of the vasculature with saline before removal of the ligated nerve did not alter the pattern of increased transferrin concentrations at the proximal ligatures, again suggesting that local vascular stasis was not involved. Finally, a slow-release implant of colchitine placed on the sciatic nerve near the lumbosacral plexus abolished transferrin accumulation at the ligatures (Fig. 3), indicating the importance of axonal microtubules in the process. Table 2 shows the results of the axonal transport studies of transferrin in regenerating sciatic nerves at the three stages of limb regeneration examined. Despite the very large increase in the overall transferrin concentration during the course of limb regeneration, the pattern of its accumulation proximal but not distal to the ligatures was similar at all stages. The portion of neural transferrin undergoing axonal transport (the mobile fraction) did not change significantly during this period, although the mean value was slightly lower at the digit stage of regeneration. The rate of anterograde transport of transferrin was also similar in the regenerating nerves at all three stages of limb regeneration (Table 2). This averaged 76 mm/day or 3.2 mm/hr, indicating that transferrin is carried in the fast vesicular component of axonal transport. The rate was the same in the contralateral intact nerves (not shown), an observation consistent with studies of fast axonal transport during nerve regeneration in mammals (McQuarrie, 1989). Transfer&

is Released at Regenerating

Ends of Axons

To examine release of transferrin from regenerating axons, sciatic nerves were dissected from adult axolotls 2-4 weeks after crush injury (not amputation) and placed in multicompartmental organ culture. One day later concentrations of transferrin were determined in culture medium from the “proximal” compartment which surrounded the ganglia and from the “distal” compartment which contained growth cones of regenerating axons. Results with six cultured nerves are shown in Fig. 4. In every case there was greater release of transferrin from the regenerating nerve distally than from the ganglia, so that the mean transferrin concentration in the distal chamber was twice that in the

396

DEVELOPMENTAL BIOLOGY

VOLUME 147,1991

FIG. 1. Myelinated fibers from axolotl sciatic nerve, teased wholemount preparation. (A) Phase contrast micrograph showing one axon (arrow) from which the myelin sheath has been stripped. (B) Same field as in (A) showing indirect immunofluorescence after treatment with anti-transferrin antiserum. Schwann sheaths and axon are both labeled. Control preparations treated with preimmune serum were completely devoid of fluorescence (not shown). (C) A bundle of myelinated axons, showing nodes of Ranvier (arrows), after indirect anti-transferrin immunofluorescent staining. (D) A single fiber with myelin sheath after anti-transferrin staining. (E) A single fiber lacking its myelin sheath after anti-transferrin staining. (F) A single fiber partially covered by a fragment of myelin sheath, shown with phase contrast (top) and fluorescent microscopy after indirect immunotluorescent staining with anti-neurofilament antibody (bottom). This antibody was used to confirm the identity of the bare axons and, as shown, was found to label only exposed regions of axons. All x500 magnification (bar = 25 pm).

chamber containing the ganglia. Cultures maintained for 3 days with media sampled every 12 hr revealed that release was linear through 24 hr, then leveled off (data not shown).

Although in vitro preparations of injured, regenerating amphibian sciatic nerves are a standard method for investigating fast axonal transport and release of proteins during neuronal growth (Hines and Garwood,

Transfer&

KIFFYEYER, TOMUSK, AND MESCHER

SEGMENTS PROXIMAL

DISTAL

FIG. 2. Transferrin concentrations in segments of regenerating sciatic nerves 3 hr after nerve ligation, expressed as percentages of the mean concentration between the ligatures. Locations of the ligatures are indicated by the Xs. Each bar indicates the mean f SE of 11 segments. Nerves were from adult axolotl hindlimbs regenerating after amputation above the ankle. Concentrations indicated by asterisk were significantly higher (P < 0.05) than the average transferrin concentration in the nerve.

1977; Tedeschi and Wilson, 1987; Snyder, 1988), it is particularly important with a plasma protein to show that release is associated with axonal transport and not simply passive diffusion from interstitial tissue space. In our preparations a greater volume of tissue was present in the proximal ganglionic chamber, suggesting that the twofold higher concentration of transferrin found in the distal compartment was not due to diffusion from the tissue. To test this point further, colchicine was added to the compartment containing the ganglia in order to block axonal transport. In every case this was found to prevent accumulation of transferrin in the medium of the distal chamber surrounding the growth cones (Fig. 5), indicating that this accumulation is dependent on anterograde transport in the axons. Axotomy Reduces the Transferrin Regeneration Blastemas

Concentration

in Regenerating Nerves

397

the distal region of normal midbud stage blastemas from larval axolotl forelimbs 6 days after amputation and in distal regions of the contralateral limb stumps denervated at the time of amputation or at different times postamputation. The concentration varied from animal to animal, but the content in the denervated tissues was always approximately half that determined for contralateral control tissues which were innervated and growing (Fig. 6). Loss of transferrin from distal tissues following axotomy was rapid: in limb stumps denervated on Day 5 postamputation, only 1 day before sampling, the transferrin concentrations were already reduced to the same extent as stumps denervated on Day 0 or Day 3. Since rapidly dividing cells take up increased amounts of transferrin, it was important to confirm that the reduced level of this factor found in the distal part of denervated limb stumps was not an indirect effect due to the absence or reduction in cell proliferation. Cell proliferation and regeneration were inhibited by unilateral X irradiation of limbs in another group of larvae. Six days after bilateral amputation the irradiated limbs showed no signs of regeneration but had concentrations of transferrin as high as those in the contralateral blastemas (Fig. 6). The reduced transferrin content in distal regions of the blastema following denervation is therefore most likely due to loss of this protein from the severed axons and its removal from the blastema. The rapidity with which the decrease occurred is consistent with the observed velocity of anterograde axonal transport (Table 2). Transferrin was never found to be completely depleted in distal tissues after denervation, which is consistent with a second source of this factor not affected by axotomy, most likely the capillaries present proximally.

in

DISCUSSION

Transferrin released distally from growing axons in viva might contribute to the local concentration of this protein. Transferrin concentrations were measured in

We have demonstrated a strong increase in the concentration of transferrin in regenerating axolotl sciatic nerve, a result that has implications for both nerve regeneration and limb regeneration. As others have re-

TABLE 2 AXONALTRANSPORT OFTRANSFERRINATTHREESTAGESOFLIMBREGENERATION Transferrin

in sciatic nerve

Average concentration (rig/mm) Concentration proximal to ligature (rig/mm) Concentration distal to ligature (rig/mm) Mobile fraction Anterograde transport rate (mm/day) Note. Mean f SE (n) is given for each value.

Early bud 28.1 -t 54.3 f 10.2 f 0.41 + 74.7 +

4.0 (4) 16.9 (4) 3.6 (4) 0.15 (4) 33.9 (4)

Palette 62.7 f 126.8 f 26.1 f 0.39 f 75.8 +

11.4 (4) 47.7 (4) 13.1 (4) 0.18 (3) 43.8 (3)

Digital outgrowth 213.7 f 49.4 (3) 354.4 + 73.2 (3) 0.22 f 0.10 (3) 76.9 f 27.5 (3)

398

DEVELOPMENTAL BIOLOGY

1

2*3

VOLUME 147, 1931

4 SEGMENTS

PROXIMAL

distal

proximal fcolchicine

DISTAL

FIG. 3. Transferrin concentrations in nerve segments like those in Fig. 2, but following local application of colchicine to the nerve distal to the lumbar plexus. Concentrations are expressed as percentages of the mean concentration between the ligatures (Xs). Each bar indicates the mean f SE of four segments. Accumulation of transferrin proximal to the ligatures is abolished, indicating its dependence on axonal transport.

FIG. 5. Transferrin concentrations in medium of culture chambers containing proximal and distal regions of regenerating nerves like those of Fig. 4, but with the addition of colchicine to medium of the proximal chamber. Colchicine in the medium surrounding the ganglia prevented release and accumulation of transferrin in the distal chamber, indicating the importance of axonal transport for this accumulation.

ported with sciatic nerve from chicken (Oh et aZ., 1981) and rats (Lin et al, 1990), peripheral nerve transferrin in the axolotl is distributed in axons and Schwann cells. Axotomy or crush injury to rat sciatic nerve stimulates a massive increase in expression of transferrin receptors on Schwann cells (Raivich et ab, 1990a) and a lofold increase in endoneural uptake of radiolabeled iron from the interstitial fluid (Raivich et aZ., 1990b). With rat facial nerve it has been shown that axotomy also rapidly induces transferrin receptors and iron uptake in neurons themselves (Graeber et ak, 1989). Cultured neurons in a serum and glia-free environment require transferrin for survival and neuritic outgrowth (Skaper

et al, 1982; Aizenman

et al., 1986; Aizenman and de Vellis, 1987), perhaps reflecting the need for iron during mitochondrial biogenesis in axons (Dion et ab, 1988). This requirement supports the view that the increased uptake of transferrin and iron by peripheral nerves after axotomy is physiologically important for axonal regeneration in viva. The highest concentration of transferrin in sciatic nerves occurred during the digit stage of limb regenera-

. 350

--

300

-. .

zoo-

.

150~-

.

.

/-

. 250 -loo-50 --

I

0’

/*

ti/ *A

.P

J

proximal

Control distal after

FIG. 4. Transferrin concentrations in medium of culture chambers containing proximal and distal regions of regenerating adult axolotl sciatic nerves maintained in organ culture for 24 hr as described under Materials and Methods. Each line represents the results with one nerve. The average concentration in the distal chambers, which contained the regenerating ends of axons injured by crush 2-4 weeks earlier, was approximately twice that in the proximal chambers containing the dorsal root ganglia.

6 days amputation

DoyO

WJ

Day 5 X-irradioted 6 days alter amputation

Denervation times after amputation

FIG. 6. Transferrin concentrations in distal tissues of denervated or X-irradiated larval forelimb stumps 6 days after amputation, given as percentages of the concentrations in the distal region of contralateral g-day (midbud) blastemas. Each bar indicates mean + SE of six limbs. Denervation at any time after amputation reduced the transferrin concentration in distal tissue by approximately 50%, but X irradiation had no effect.

KIFFMEYER,TOMUSK,ANDMESCHER

tion, when axonal elongation is nearly complete and redifferentiation is underway. Based on the work of others, the large amount of the factor in the nerve at this time may correlate with roles for iron during synaptogenesis (Mollgard et al, 1984) and myelination (Lin et al., 1990). Why transferrin also accumulates in contralateral sciatic nerves not directly affected by amputation is not clear, but similar bilateral reactions to unilateral nerve injury are well known (Barr and Hamilton, 1948; Tweedle, 1971; Wells and Vaidya, 1989). The elevated concentration of transferrin in growing nerves is of special interest in light of the enhanced ability of such nerves to promote proliferation in the regeneration blastema. Boilly and Albert (1988) recently demonstrated higher growth-promoting activity for cultured blastema cells in spinal cord extracts from axolotls with regenerating limbs than in similar extracts from control animals. Moreover, a previous injury to axons causing them to regenerate significantly increases their ability to promote proliferation of blastema cells in vitro (Boilly and Bauduin, 1988) and growth of a new limb in viva (Maier et al., 1984). Our results suggest that the improved activity of the growing nerves in these studies may be due to higher concentrations of transferrin. The results of the ligature studies clearly demonstrate that transferrin is carried in the fast component of axonal transport, which has long been considered the most likely mechanism for delivery of trophic factor(s) in peripheral nerve (Wallace, 1981). The anterograde velocity we measured, 76 mm/day, falls within the range of rates reported by others for fast axonal transport in amphibian sciatic nerves (Stone et al, 1978; Hines and Garwood, 1977; Edstrom and Mattsson, 1972; Partlow et al., 1972). Transferrin is the first substance required for growth of blastema cells that has been shown to be carried by fast axonal transport. For axonally transported transferrin to exert a trophic effect on blastema growth, it would have to be released distally from axonal growth cones. We have shown that transferrin is released from regenerating ends of axolotl sciatic nerve using an organ culture technique developed by others to study secretion of fasttransported material (Stone et a.!., 1978; Tedeschi and Wilson, 1986,1987; Snyder, 1988). Control cultures with colchicine confirmed that transferrin released distally is carried by fast axonal transport. Our observation that the amount of transported transferrin increases during amphibian limb regeneration (Table 2) is consistent with results of Tedeschi and Wilson (1986, 1987), who found that distal release of several unidentified, fasttransported glycoproteins increased during the course of nerve regeneration.

Transfwrin

in Regenerating Nerves

399

Released from a cell after delivery of iron, apotransferrin can bind locally available iron for another receptor-mediated uptake cycle with the same or a different cell (reviewed by May and Cuatrecasas, 1985; Huebers and Finch, 1987). A process involving axonal transport, release, and recycling of a factor required for both axonal growth and cell cycling closely resembles a proposal of Singer (1964,1978) regarding the trophic effect of nerves. He argued that the primary target for the neural factor(s) is most likely the axons themselves, where it is required for neuritic elongation, plasticity, and maintenance. Moreover, upon discharge of the factor from neurons, the same properties that made it important for the axons would enable it to stimulate metabolism and growth of surrounding cells (Singer, 1964, 1978). The physiological characteristics of transferrin, particularly its reuse after exocytosis, and its importance for both neurite growth and cell proliferation fit these predictions very well. Our finding that denervation lowers the transferrin concentration in distal limb stump tissues provides further support for the hypothesis that neural delivery of this factor is important for limb regeneration. As expected for factors delivered by fast axonal transport, reduction of the transferrin concentration occurred within 24 hr of axotomy. However, even by 5 days after denervation the transferrin content never fell to less than half that in contralateral control blastemas. A basal level of transferrin in distal limb stump tissue is evidently contributed from nonaxonal source(s), perhaps from blood vessels present proximally. This result supports the argument of Brockes (1987) that there must be two “systems” of blastema growth control, one dependent on axons and one based possibly on circulating factors, since dedifferentiated cells are not all affected equally by denervation. In denervated nonregenerating limb stumps of axolotl larvae like those assayed here for transferrin, Barger and Tassava (1985) found that over 50% of the cells enter the S period between Days 4 and 6 postamputation despite the absence of nerves. Such observations are consistent with the concept that the neural effect is based on axonal release of a permissive growth factor that is also available from the vasculature, but not in a quantity sufficient to support blastema formation. Results of the present in viva experiments strengthen the following hypothesis, which is discussed in more detail elsewhere (Mescher and Munaim, 1988). Transferrin is transported axonally to deliver iron needed for axonal regeneration and apotransferrin released from the axons binds additional iron locally. In the distal dedifferentiated tissue of a regenerating limb at the early bud stage, axonally derived transferrin supplements that available by diffusion from capillaries in the stump

400

DEVELOPMENTAL BIOLOGY

and together these two sources provide a concentration of this factor adequate to sustain cycling of mesenchyma1 cells. When paracrine release from the axons is cut off by denervation, the vascular supply alone cannot maintain a sufficient level of transferrin and cells cannot proliferate adequately to form a blastema. Is this hypothesis consistent with what is known regarding the neural effect on limb regeneration? It has been appreciated for several years that the trophic influence of axons on blastema growth seems to promote general synthetic and proliferative activity of mesenchyma1 cells by a rather nonspecific effect on metabolic rate (reviewed by Globus, 1978; Singer, 1978; Wallace, 1981). Nerves are not needed for the initiation of cycling in dedifferentiating cells, but are required for this activity to continue (Tassava and Olsen, 1985). In a recent analysis of blastema growth control, Muneoka et al. (1989) concluded that the important agents from nerves are not regulatory factors determining the onset and termination of growth, but are instead requirements necessary for cells to sustain growth. The concept that the trophic effect of nerves is based on release of permissive factors like transferrin needed for cell cycle progression, rather than mitogenic growth factors that signal cell cycle entry, is consistent with these earlier ideas. Availability of a permissive growth factor from either axons or capillaries would explain the gradual independence from nerves regenerating limbs exhibit once mesenchymal cells have accumulated (reviewed by Singer, 1952). Mitotic activity is diminished after denervation at any stage of regeneration, but less profoundly in later stages than in early ones (Singer, 1952; Kintner and Brockes, 1985). Inspection of the descriptive literature on limb regeneration reveals that the gradual emancipation from complete dependence on nerves for cell proliferation is concomitant with ingrowth of capillaries into the blastema from vessels in the stump. Vascular marking studies with adult newts by Peadon and Singer (1966) indicated that the early, nerve-dependent blastema is essentially avascular and that the most rapidly growing regions in the later regenerate remain those most poorly vascularized. These observations were confirmed by the histological analysis of Iten and Bryant (1973) and are similar to the results when vascular development was examined in regenerating limbs of larval salamanders (Revardel and Chapron, 1975; Smith and Wolpert, 1975). These studies all demonstrate that capillaries begin to sprout from proximal regions of the stump into the growing blastema itself only during the early bud stage of regeneration. The importance of new capillaries to supply nutrients and other factors for continued division in an accumulation of cells is well-known from work correlating avascular microenvironments, diffusion distances, and proliferative quiescence in tu-

VOLUME 147.1991

mor-like models (reviewed by Sutherland, 1988). If capillaries entering the blastema can deliver growth-promoting factors similar to those released by axons, the neural dependence for most proliferation that exists before vascularization would be relieved. In conclusion, data presented here show that the permissive growth factor transferrin is accumulated, axonally transported, and released distally by regenerating amphibian sciatic nerves. The loss of this factor from early bud stage blastemas following axotomy suggests that neurally derived transferrin may be important in sustaining proliferation during the nerve-dependent, avascular phase of limb regeneration. Other nutrient transport factors, such as apolipoproteins (Boyles et cd, 1990), also accumulate in regenerating peripheral nerve and could contribute to the trophic effect of nerve if similarly released from axons. Paracrine release of transferrin and other factors from axons may be particularly important for cells with insufficient access to proteins from plasma. The molecular basis of neural development in the regenerating limb is highly complex and the axonal influence on proliferation of surrounding cells is likely to involve many factors, acting in combination with agents from the wound epithelium and from cells mediating the inflammatory response to promote blastema formation. We thank Sidney Ochs for assistance during the initial axonal transport studies and Gene Weinberg for helpful discussions. The work was supported in part by grants from the U.S. Army Medical Research and Development Command (DAMDl?-87-C-7098 and DAMD17-91-Z1002) to A.L.M. REFERENCES AIZENMAN, Y., and DE VELLIS J. (1987). Brain neurons develop in a serum and glial free environment: effects of transferrin, insulin, insulin-like growth factor-I and thyroid hormone on neuronal survival, growth and differentiation. Brain Res. 406,32-42. AIZENMAN, Y., WEICHSEL, M. E., and DE VELLIS, J. (1986). Changes in insulin and transferrin requirements of pure brain neuronal cultures during embryonic development. Proc. Natl Acad Sci. USA 83, 2263-2266. ALBERT, P., and BOILLY, B. (1988). Effect of transferrin on amphibian limb regeneration: A blastema cell culture study. Wilhelm Roux Archin Lkv. Biol. 197,193-196. BARGER, P. M., and TASSAVA, R. A. (1985). Kinetics of cell cycle entry in innervated and denervated forelimb stumps of larval Am&stoma. J. Exp. Zoo1 233,151-154. BARNES, D. and SATO, G. (1980). Methods for growth of cultured cells in serum-free medium. Anal Biochem. 102,255~270. BARR, J., and HAMILTON, J. (1948). A quantitative study of certain morphological changes in spinal motor neurons during axon regeneration. J. Camp. Neural. 89.93-121. BEACH, R. L., POPIELA, H. and FESTOFF,B. W. (1983). The identification of neurotrophie factor as a transferrin. FEBS Mt. 156, 151156. BISBY, M. A. (1982). Ligature techniques. In “Axoplasmic Transport” (D. G. Weiss, Ed.), pp. 193-199. Springer-Verlag, Berlin.

KIFFMEYER, TOMUSK, AND MESCHER BOILLY, B., and ALBERT, P. (1988). Blastema cell proliferation in vitro: Effects of limb amputation on the mitogenic activity of spinal cord extracts. Biol Cell 62,183-187. BOILLY, B., and BAUDUIN, B. (1988). Production in vitro by spinal cord of growth factor(s) acting on newt limb regeneration: Influence of regeneration of the nerve fibers. Dev. Brain Res. 38,155-160. BOWMAN, B. H., YANG, F. and ADRIAN, G. S. (1988). Transferrin: Evolution and genetic regulation of expression. A&J. Genet. 25, l-38. BOYLES, J. K., NOTTERPEK, L. M., and ANDERSON, L. J. (1990). Accumulation of apolipoproteins in the regenerating and remyelinating mammalian peripheral nerve. J. Biol. Chem. 265,17,805-1’7,815. BROCKES,J. P. (1987). The nerve dependence of amphibian limb regeneration. J. Exp. Biol. 132,79-91. CARLONE, R. M., and RATHBONE, M. P. (1985). Partial purification of a low-molecular-weight growth factor from chicken brain. J. Cell Physial

124,467-473.

CHOO, A. F., LOGAN, D. M. and RATHBONE, M. P. (1978). Nerve trophic effects: An in vitro assay for factors involved in regulation of protein synthesis in regenerating amphibian limbs. J. Exp. Zool 206, 347-354. CLARK, R. A. F., and HENSON, P. M. (Eds.) (1988). “The Molecular and Cellular Biology of Wound Repair.” Plenum, New York. DAVIS, R. J., and CZECH,M. P. (1986). Regulation of transferrin receptor expression at the cell surface by insulin-like growth factors, epidermal growth factor and platelet-derived growth factor. EMBO J. 5,653-658. DE JONG, G., VAN DIJK, J. P., and VAN EIJK, H. G. (1990). The biology of transferrin. Clin Chim. Ada 190, l-46. DION, T. L., MARKELONIS, G. J., OH, T. H., BREGMAN, B. S., PUGH, M. A., HOBBS, S. L., and KIM, Y. C. (1988). Immunocytochemical localization of transferrin and mitochondrial malate dehydrogenase in the developing nervous system of the rat. Dev. Neurosci. 10, 152-164. EDSTROM, A., and MA~TSSON, H. (1972). Fast axonal transport in vitro in the sciatic system of the frog. J. Neurochem. 19,205-221. GLOBUS, M. (1978). Neurotrophic contribution to a proposed tripartite control of the mitotic cycle in the regeneration blastema of the newt, Notophthalmus (Triturus) viridescens. Am. Zoo1 18,855~868. GU)BUS, M., and VETHAMANY-GLOBUS, S. (1985). In vitro studies of controlling factors in newt limb regeneration. In “Regulation of Vertebrate Limb Regeneration” (R. E. Sicard, Ed.), pp. 106-127. Oxford Univ. Press, New York. GOLDHAMMER, D. J., and TASSAVA, R. A. (1987). An analysis of proliferative activity in innervated and denervated forelimb regenerates of the newt, Notophthalmus viridescens. Development 100,619-628. GRAEBER, M. B., RAIVICH, G., and KREUTZBERG, G. W. (1989). Increase of transferrin receptors and iron uptake in regenerating motor neurons. J. Neurosci. Res. 23,342-345. HINES, J. F., and GARWOOD, M. M. (1977). Release of protein from axons during rapid axonal transport: an in vitro preparation. Brain Res. 125,141-148. HUEBERS, H. A., and FINCH, C. A. (1987). The physiology of transferrin and transferrin receptors. Physiol. Rev. 67,520-582. ITEN, L. E., and BRYANT, S. V. (1973). Forelimb regeneration from different levels of amputation in the newt, Notophthalmus vimdescens: Length, rate and stages. Wilhelm Roux Arch. Dev. Biol. 173, 263-282. KINTNER, C. R. and BROCKES, J. P. (1985). Monoclonal antibodies to cells of a regenerating limb. J. Embryo1 Exp. Morph01 89,37-55. LIN, H. H., SNYDER, B. S., and CONNOR,J. R. (1990). Transferrin expression in myelinated and non-myelinated peripheral nerves. Brain

Res. 526,217-220.

MAIER, C. E., MCQUARRIE, I. G., and SINGER, M. (1984). A nerve-condi-

Transf&rrin

in Regenerating

Nerves

401

tioning lesion accelerates limb regeneration in the newt. J. E~J. Zool. 232,181-186. MARKELOMS, G. J., BRADSHAW, R. A., OH, T. H., JOHNSON,J. L., and BATES, 0. J. (1982). Sciatin is a transferrin-like polypeptide. J. Ne-urochem 39,315-20. MAY, W. S., and CUATRECASAS, P. (1985). Transferrin receptor: Its biological significance. J. Me&r. Biol. 88,205-215. MCQUARRIE, I. G. (1989). Axonal transport and the regenerating nerve. In “Neural Regeneration and Transplantation,” (F. J. Seil, Ed.), pp. 29-42. Liss, New York. MEEK, J., and ADAMSON, E. D. (1985). Transferrin in foetal and adult mouse tissues: Synthesis, storage and secretion. J. Embryo1 Exp. Mwphol.

86,205-2X%

MESCHER, A. L., and LOH, J. J. (1981). Newt forelimb blastemas in vitro: Cellular response to explantation and effects of various growth-promoting substances. J. Exp. Zool. 216,235-245. MESCHER, A. L., and MUNAIM, S. I. (1984). “Trophic” effect of transferrin on amphibian limb regeneration blastemas. J. Eqn. Zool. 230, 485-490.

MESCHER, A. L., and MUNAIM, S. I. (1988). Transferrin and the growthpromoting effect of nerves. Intl Rev. Cytol. 110, l-26. MOLLGARD, K., REYNOLDS,M., JACOBSEN, N., DZIEGIELEWSKA, K., and SAUNDER, K. (1984). Differential immunocytochemical staining for fetuin and transferrin in the developing cortical plate. J. Neurocyto1. 13,497-502. MUNAIM, S. I., and MESCHER, A. L. (1986). Transferrin and the trophic effect of neural tissue on amphibian limb regeneration blastemas. Dev. Biol. 116,138-142. MUNEOKA, K., BRYANT, S. V., and GARDINER, D. M. (1989). Growth control in limb regeneration. In “Developmental Biology of the Axolotl” (J. B. Armstrong and G. M. Malacinski, Eds.), pp. 143-156. Oxford Univ. Press, New York. OH, T. H., SOFIA, C. A., KIM, Y. C., CARROLL, C., KIM, H. H., MARKELONIS, G. J., and REYER, P. J. (1981). Sciatin: Immunocytochemical localization of a myotrophic protein in chicken neural tissues. J. Histochem. Cytochem 29,1205-1212. PARDEE, A. B., CHERINGTON, P. V., and MEDRANO, E. E. (1981). On deciding which factors regulate cell growth. In “Control of Cellular Division and Development, Part A” (D. Cunningham et al, Eds.), pp. 495-502. Liss, New York. PARTLOW, L. M., Ross, C. D., MOTWANI, R., and MCDOUGAL, D. B. (1972). Transport of axonal enzymes in surviving segments of frog sciatic nerve. J. Gen. Physiol. 60,388-405. PEADON, A. M., and SINGER, M. (1966). The blood vessels of the regenerating limb of the adult newt, Triturus. J Morphol. 118,79-90. PEREZ-INFANTE, V., and MATHER, J. P. (1982). The role of transferrin in the growth of testicular cell lines in serum-free medium. Exp. Cell Res. 142,325-332. RAIVICH, G., HELLWEG, R., GRAEBER, M. B., and KREUTZBERG, G. W. (1990a). The expression of growth factor receptors during nerve regeneration. Restor. New01 Neurosd 1,217-223. RAIVICH, G., GRAEBER, M. B., GEHRMANN, J., and KREUTZBERG, G. W. (1990b). Transferrin receptor expression in the injured and regenerating rat sciatic nerve. Sot. Neurosci. Abstr. 16,163. RASOOL, C. G., SCHWARTZ, A. L., BOLLINGER, J. A., REICHLIN, S., and BRADLEY, W. G. (1981). Immunoreactive somatostatin distribution and axoplasmic transport in rat peripheral nerve. Endocrinology 108,996-1001.

REICHARD, P. (1988). Interactions between deoxyribonucleotide and DNA synthesis. Annu. Rev. Biochem. 57,349-374. REVARDEL, J.-L., and CHAPRON, C. (1975). Influence de la vascularisation sur regeneration des membres chez les larves d’urodeles. Nouvelle interpretation du role du systeme nerveux. C, R. Acad. Sci Paris 280(D), 1409-1411.

402

DEVELOPMENTAL BIOLOGY

ROBERTS, M. R., MISKIMINS, W. K., and RUDDLE, F. H. (1989). Nuclear proteins TREFl and TREFZ bind to the transcriptional control element of the transferrin receptor gene and appear to be associated as a heterodimer. Cell Reg. 1,151-164. SCADDING, S. R. (1988). Treatment of brachial nerves with colchicine inhibits limb regeneration in the newt Notophthalmus viridescens. J. Exp. Zool. 247,56-61. SELIGMAN, P. (1983). Structure and function of the transferrin receptor. Prog. Hematol. 13,131-147. SINGER, M. (1952). The influence of the nerve in regeneration of the amphibian extremity. Quar. Rev. Biol. 27, 169-200. SINGER, M. (1964). The trophic quality of the neuron: some theoretical considerations. Prog. Brain. Res. 13,228-323. SINGER, M. (1978). On the nature of the neurotrophic phenomenon in urodele limb regeneration. Am. Zool. l&829-41. SKAPER, S. D., SELAK, I., and VARON, S. (1982). Molecular requirements for survival of cultured avian and rodent dorsal root ganglionic neurons responding to different trophic factors. J. Neurosti Res. 8, 251-261. SMITH, A. R., and WOLPERT, L. (1975). Nerves and angiogenesis in amphibian limb regeneration. Nature 257,224-225. SNYDER, R. E. (1988). The release of axonally transported material from an in vitro amphibian sciatic nerve preparation. J. Neurobiol. 19,283-292. STONE, G. C., WILSON, D. L., and HALL, M. E. (1978). Two-dimensional gel electrophoresis of proteins in rapid axoplasmic transport. Brain Res. 144,287-302. SUTHERLAND, R. M. (1988). Cell and environment interactions in tumor microregions: The multicell spheroid model. Science 240,177184.

VOLUME 147, I991

TANK, P. W., CARLSON, B. M., and CONNELLY, T. G. (1976). A staging system for forelimb regeneration in the axolotl, Ambystoma mexicanum J. Morph& 150,117-128. TASSAVA, R. A., and MESCHER, A. L. (1975). The roles of injury, nerves, and the wound epidermis during the initiation of amphibian limb regeneration. DiJTerentiation 4,23-24. TASSAVA, R. A., and OLSEN, C. L. (1985). Neurotrophic influence on cellular proliferation in urodele limb regeneration: In vivo experiments. In “Regulation of Vertebrate Limb Regeneration” (R. E. Sicard, Ed.), Oxford Univ. Press, pp. 81-92. New York. TEDESCHI, B., and WILSON, D. L. (1986). The destination and modification of fast-transported proteins in normal and regrowing nerve. In “Axoplasmic Transport” (Z. Iqbal, Ed.), pp. 263-273. CRC Press, Boca Raton, FL. TEDESCHI, B., and WILSON, D. L. (1987). Subsets of axonally transported and periaxonal polypeptides are released from regenerating nerve. J. Neurochem. 48,463-469. TETZLAFF, W., ZWIERS, H., LEDERIS, K., CASSAR, L., and BISBY, M. A. (1989). Axonal transport and localization of B-50/GAP-43-like immunoreactivity in regenerating sciatic and facial nerves of the rat. J. Neurosti 9,1303-1313. TWEEDLE, C. (1971). Transneuronal effects of amphibian limb regeneration. J. Exp. Zoul. 177,13-30. WALLACE, H. (1981). Vertebrate Limb Regeneration. Wiley, New York. WELLS, M. R., and VAIDYA, U. (1989). Morphological alterations in dorsal root ganglion neurons after peripheral axon injury: Association with changes in metabolism. Exp. Neural 104,32-38. WERNER, P. A. M., GALBRAITH, R. M., and ARNAIJD, P. (1983). DEAEAffi-Gel Blue chromatography of human serum: Use for purification of native transferrin. Arch. B&hem. Biophys. 226,393-398.