Transforming Growth Factor-/ l Regulates Axon/Schwann Cell Interactions

We have investigated the potential regulatory role of TGF-B in the interactions of neurons and Schwann cells using an in vitro myelinating system. Purified populations of neurons and Schwann cells, grown alone or in coculture, secrete readily detectable levels of the three mammalian isoforms of TGF-B; in each case, virtually all of the TGF-/~ activity detected is latent. Expression of TGF-BI, a major isoform produced by Schwann cells, is specifically and significandy downregulated as a result of axon/Schwann cell interactions. Treatment of Schwann cells or Schwann cell/neuron cocultures with TGF-~I, in turn, has dramatic effects on proliferation and differentiation. In the case of purified Schwann cells, treatment with TOF-/31 increases their proliferation, and it promotes a preor nonmyelinating Schwann cell phenotype characterized by increased NCAM expression, decreased NGF receptor expression, inhibition of the forskolin-mediated induction of the myelin protein P0, and induction of the Schwann cell transcription factor suppressed cAMPinducible POU protein. Addition of TGF-/31 to the cocultures inhibits many of the effects of the axon on Schwann cells, antagonizing the proliferation induced by contact with neurons, and, strikingly, blocking myelination. Ultrastructural analysis of the treated cultures confirmed the complete inhibition of myelination and revealed only rudimentary ensheathment of axons. Associated defects of the Schwann cell basal lamina and reduced expression of laminin were also detected. These effects of TGF-/?I on Schwann cell differentiation are likely to be direct effects on the Schwann cells themselves which express high levels of TGF-/~I receptors when cocultured with neurons. The regulated expression of TGF-~ and its effects on Schwann cells suggest that it may be an important autocrine and paracrine mediator of neuron/Schwann cell interactions. During development, TGF-~I could serve as an inhibitor of Schwann cell proliferation and myelination, whereas after peripheral nerve injury, it may promote the transition of Schwann cells to a proliferating, nonmyelinating phenotype, and thereby enhani:e the regenerative response. p ERIPHERAL nerve development progresses through a series of distinct stages that reflect complex and reciprocal interactions between axons and Schwann cells (Webster, 1992). Initially, nerve fibers grow out essentially free of nonneuronal cells. Subsequently, Schwann cells migrate and proliferate on the nerve fibers, progressively subdividing the nerve fascicle into smaller groups of nerve fibers. Eventually, Schwann cells either communally ensheathe multiple small nerve fibers, or they myelinate individual nerve fibers with which they have established a oneto-one relationship (Webster, 1992). In both instances, the axon/Schwann cell unit is surrounded by a basal lamina that is principally synthesized by Schwann cells when they are in Address correspondence to Dr. James Salzer, Department of Cell Biology, New York University Medical School, 550 First Avenue, New York, NY 10016 Tel.: (212)263-5358. Fax: (212)263-8139. The current address for Dr. Metz is The Picower InStitute for Medical Research, 350 Community Drive,.Manhasset, NY 11030. contact with neurons (Bunge et al., 1982). This basal lamina, in turn, is required for the appropriate function of the Schwann cell; defects in basal lamina production are associated with defects in the ensheathment and myelination of axons (Bunge et al., 1986). The different anatomic relationships that Schwann cells establish with axons correspond to distinct differentiated phenotypes. Nomnyelinating, ensheathing Schwann cells express high levels of the adhesion molecules L1 and the neural cell adhesion molecule (NCAM) ~ do not express myelin proteins, and contain a distinct set of cytoskeletal proteins. By contrast, myelinating Schwann cells express low levels of NCAM and L1, but high levels of the myelin as1. Abbreviations used in :his paper: DRG, dorsal root ganglion; MAG, myelin-associated glycoprotein; MBP, myelin basic protein; BrDU, 5-bromo2'deoxyuridine; NCAM, neural cell adhesion molecule; SCIP, suppressed cAMP-inducible POU protein; MLEC, mink lung epithelial cells; PAI, plasminogen activator inhibitor. © The Rockefeller University Press, 0021-9525/95/04/443/16 $2.00 The Journal of Cell Biology, Volume 129, Number 2, April 1995 443-458 443 on A uust 4, 2017 jcb.rress.org D ow nladed fom sociated glyeoprotein (MAG) and several structural components of the myelin sheath, including myelin basic protein (MBP) and P0 (reviewed in Jessen and Mirsky, 1991; Salzer, 1995). Like proliferation and basal lamina formation, the differentiation of Schwann cells is regulated by the axon as their ensheathment fate is specified by the type of axon with which they are associated (Aguayo et al., 1976; Weinberg and Spencer, 1976). After injury, there are dramatic changes in peripheral nerves distal to the site of the lesion (reviewed in Fawcett and Keynes, 1990). Expression of myelin proteins rapidly declines (Lernke and Chao, 1988) and Schwann cells reexpress NCAM and L1 (Martini and Schachner, 1988). Schwann cells also undergo a wave of proliferation at the site of injury in all nerves and in the distal stump of heavily myelinated nerves (reviewed briefly in Salzer and Bunge, 1980). Concomitantly, macrophages invade the distal stump and, together with resident Schwann cells, they break down and clear the degenerating myelin (Stoll et al., 1989). Finally, if the nerve has not been permanently transected, new nerve fibers sprout from the proximal end of the injured nerve and grow into the distal stump guided by Schwann cells. The downregulation and clearance of myelin proteins and the reexpression of cell adhesion molecules by Schwann cells, together with their synthesis and release of neurotrophic factors, is thought to be critical for successful regeneration of peripheral nerves after injury (Scherer and Asbury, 1993). The molecular signals that regulate Schwann cell proliferation and differentiation during peripheral nerve development and injury are not well understood. Molecules associated with the neuronal surface are known to induce Schwann cell proliferation during development (Salzer et al., 19g0), and they are likely to correspond, at least in part, to the recently described family of neuregulins (Marchionni et al., 1993), also referred to as glial growth factor and heregulin (Peles and Yarden, 1993). It is not yet known whether the mitogenic signals that stimulate Schwann cell proliferation during nerve injury are related to those that operate during development. However, a recent study reporting that soluble mitogenic factors are released as a result of injury (Wen et al., 1994) suggests that they could be distinct. An additional potential source of mitogenic factors during nerve injury are macrophages, which invade the distal stump in high numbers and have been proposed to promote Schwann cell proliferation during injury (Beuche and Friede, 1984) via soluble factors (Baichwal et al., 1988). However, the proliferation of Schwarm cells during Wallerian degeneration in vitro, in the absence of macrophages (Salzer and Bunge, 1980), suggests that nonmacrophage-derived mitogens, potentially growth factors released by the Schwann cells themselves, are also likely to be important. Both proliferative and differentiative signals may be mediated via an increase in intracellular levels of cAMP. At low cell density, elevation of cAMP levels increases Schwann cell proliferation, whereas at high cell densities, particularly when proliferation is limited by growing cells in defined media without serum, an increase in cAMP leads to an increase in the expression of myelin proteins (Jessen and Mirsky, 1991; Morgan et al., 199I). In view of these findings, treatment of Schwann ceils with the diterpene analogue forskolin, an activator of adenylate cyclase, has been used to mimic the effects of the axon on Schwann cells, particularly inducing the myelinating phenotype. Forskolin treatment, in addition to increasing expression of myelin proteins, also dramatically elevates the expression of the transcription factor SCIP (suppressed cAMP-inducible POU protein) by Schwann ceils (Monuki et al., 1989). This protein, which was independently identified by several groups and is also called tst-1 (He et al., 1990) and Oct-6 (Suzuki et al., 1990), belongs to the POU domain family of transcription factors. It is prominently expressed by Schwann cells during periods of rapid proliferation, i.e., during development and transiently after peripheral nerve injury (Monuki et al., 1990; Scherer et al., 1994). Its expression in vivo has also been inversely correlated with the ability of Schwann cells to myelinate. Therefore, SCIP has been suggested to be a marker of, and may function in, proliferating, premyelinating Schwann ceils. Consistent with this suggestion, SCIP inhibits the expression of myelin proteins, strongly repressing the P0 promoter, and also inhibiting the expression of the p75 NGF receptor (Monuki et al., 1990). These studies suggest that there is an inverse relationship between Schwann cell proliferation and expression of the myelinating phenotype. Consistent with this notion, several Schwann cell mitogens, notably FGF and members of the TGF-fl family (Ridley et al., 1989; Davis and Stroobant, 1990; Schubert, 1992), have been reported to inhibit the expression of myelin proteins induced by forskolin (Mews and Meyer, 1993; Rogister et al., 1993; Morgan et al., 1994). These results raise the possibility that these growth factors may have a role as inhibitors of Schwann cell myelination, although it is not known under what conditions and by which cells these growth factors are released. In contrast, axons may also release soluble factor(s) that promote the myelinating phenotype, transiently inducing the expression of SCIP, for example, and leading to a several-fold increase in P0 expression (Bolin and Shooter, 1993). These findings suggest that soluble mediators might function as both positive and negative regulators of Schwann cell myelination. In this study, we have investigated the role of TGF-/3s as mediators of axon/Schwann cell interactions. We focused on the TGF-/~s because of their critical role in regulating cell growth and differentiation in many developing systems (Massagu6, 1990; Sporn and Roberts, 1992) and their mitogenic effects on Schwann cells. The "I'GF-/~ family is comprised of three mammalian isoforms termed TGF-fll, -/32, and -f13, and the TGF-/3 superfamily contains, in addition, a large number of homologous proteins (see Kingsley, 1994, for a recent review). We have principally focused on the role of TGF-ffl in this study because, as we now report, TGF-fll is a prominent isoform produced by Schwann cells, and the expression of TGF-/31 appears to be regulated by axon/Schwann cell interactions. We have found that TGF-fll, which is a mitogen for purified Schwann cells, inhibits the proliferation of Schwann cells that is normally induced by neuronal contact. In addition, TGF-fll increases the expression of markers of premyelinating Schwarm cells, notably NCAM and SCIP, and it inhibits the forskolin-induced transition to a myelinating phenotype. Consistent with these findings, TGF-fll strikingly inhibits myelination in Schwann cell/neuron cocultures and leads to associated defects in the formation of the Schwann cell basal lamina. These results indicate that TGF-/31 has profound effects on the axonal induction of Schwann cell proliferation and differentiation, and they sugThe Journal of Cell Biology, Volume 129, 1995 444 on A uust 4, 2017 jcb.rress.org D ow nladed fom gest that it could be an important inhibitor of Schwann cell proliferation and myelination during development. These studies also suggest an important role for TGF-ffl in the regenerative response that follows Wallerian degeneration in the peripheral nervous system. Materials and Methods Antibodies and Growth Factors Antibodies used in this study included anti-MBP and anti-PO polyclonal antibodies (gifts from D. Colman, Mount Sinai Medical Center, NY), a polyclonal anti-p75 NGF receptor antibody (gift from Dr. B. Hempstead, Cornell University Medical Center, NY), monoclonal antibody MAS13 to the MAG (gift from M. Schachner, Swiss Federal Institute of Technology, H6nggerberg, Ztirich), anti-Ll polyclonal antibody (gift from M. Grumet, New York University Medical School, NY), anti-NCAM polyclonal antibody (gift from U. Rutishauser, Case Western Reserve University, Cleveland, OH), anti-SCIP polyclonal antibody (gift from M. G. Rosenfeld, University of California, San Diego, CA) and anti-laminin polyclonal antibody (Sigma Chemical Co., St. Louis, MO). Neutralizing antibodies against recombinant human TGF-/~I, native porcine "I'GF-/~2, and recombinant chicken TGF-/33 were purchased from R & D Systems, Inc. (Minneapolis, MN). A monoclonal antibody that neutralizes rat TGF-~I, TGF/32, and TGF-~3 was purchased from G-enzyme Corp. (Cambridge, MA). Species-specific, affinity-purified, rhodamine-conjugated donkey anti-rabbit IgG and fluorescein-conjugated donkey anti-mouse IgG were purchased from Chemicon International Inc. (Temecula, CA); fluorescein-conjugated mouse monoclonal antibodies directed against bromodeoxyuridine (BrDU) were obtained from Boehringer Mannheim Corp. (Indianapolis, IN). Recombinant human TGF-~I was a gift from Berlex Biosciences (South San Francisco, CA), recombinant human TGF-~2 was a gift from Celtrix (Santa Clara, CA), and TGF-B3 was purchased from R & D Systems. All concentrated stocks of TGF-/3 were stored at 4°C in a solution of 5 mM HCI containing 1 mg/ml low endotoxin BSA (ICN Biomedicals, Inc., Costa Mesa, CA). ~ssue Culture Methods Cultures of primary rat Schwann cells, dorsal root ganglion (DRG) neurons, and myelinating Schwann celI/DRG neurons were established as described previously (Einheber et al., 1993). Briefly, cultures of dissociated rat embryonic day 16 DRG neurons were grown on collagen-coated 12-ram glass coverslips in a four-well dish (Nunc, Naperville, IL) and cycled with antimitotic agents in standard serum containing media to remove nonneuronal cells. The standard media consists of MEM (Whittaker Biopreducts, Inc., Walkersville, MD) supplemented with 10% FBS, 2 mM glntamine, 0.4% glucose, and 50 ng/ml 2.5S NGF (Bioproducts for Science, Inc., Indianapolis, IN)'. To establish myelinating cultures, DRG neuron cultures were seeded with 200,000 Schwann cells in standard media. Based on cell morphology, these Schwann cell preparations contained fewer than 0.1% fibroblasts. On the next day, the standard media was replaced with N2 defined media (5 mg/ml insulin, 10 mg/ml transferrin, 20 nM progesterone, 100 mM putrescine, 30 nM selenium, and 2 mM glutamine in a 1:1 mixture of DME and Ham's F-12 supplemented with 2.5S NGF). In this media, Schwann cells in contact with neurites proliferate in response to a neuronal mitogen, but they do not assemble a basal lamina or myelinate (Moya et ai., 1980). The cultures were maintained in N2 media for 3 d to allow the Schwann cells to populate the neurites. To initiate basal lamina formation and myelination, the cultures were fed the standard media supplemented with 50 mg/ml ascorbic acid. Determination of TGF-{3 Activity Concentrations of total and active TGF-15 present in serum-free culture supernatants were measured using the plasminogen activator inhibitor-I promoter luciferase (PAI/L) assay as described (Abe et al., 1994). In this assay, serum-free culture conditioned superuatants are incubated with mink lung epithelial cells (MLEC) stably transfected with an expression construct containing a portion of the TGF-/~-inducible plasminogen activator inhibitor-I (PAI-1) promoter fused to the firefly luciferase reporter gene. Exposure of the transfected MLEC to TGF-/3 results in a dose-dependent increase in luciferase activity in lysates of the cells, as measured with a luminometer. 12 separate coverslips, each containing dissociated DRG neurons seeded with 200,000 Schwann cells, were maintained on N2 media for 3 d. Six of these cultures were switched to standard media (nonmyelinating cultures) and six to standard media supplemented with ascorbic acid (myelinating cultures). Parallel cultures of neurons alone and Schwann cells alone (400,000 Schwann cells/collagen coated coverslip) were maintained as described for the nonmyelinating cultures. (This number of Schwann cells represents an estimate of the minimum number of cells actually present in the cocultures based on cell counts from random fields). The cultures were fed their respective media every 2 or 3 d for a total of 8 d. The cultures were washed three times with N2 media, and they were incubated in 0.2 ml of N2 media for an additional 2 d. The conditioned N2 media from each group of cultures was collected, pooled, and spun 10 rain at 4°C in a tabletop centrifuge to remove cellular debris. To determine the levels of total TGF-B (active and latent), conditioned media were heated for 12 rain at 80°C to activate latent TGF-B. Active TGF-/3 levels were determined from unheated, undiluted conditioned media. Other samples were diluted to 20% with N2 media and added to the PAI/L-transfected MLEC in 96-well plates. To determine the total amount of PAL1 promoter activity specifically induced by TGF-/3 in the samples, a neutralizing anti-TGF-/31,2,3 monoclonal antibody (20/~g/ml) was added to the diluted conditioned media and incubated with the transfected MLEC. The decrease in the amount of luciferase expressed in the presence of this antibody (,x,75 % of the total luciferase in each case) was used to calculate the amount of active and total (active plus latent) TGF-/3 in the conditioned media. Similarly, the amount of PAI-1 promoter activity induced by the individual TGF-/~ isoforms was determined by the addition of TGF-/31, -if2, or -/33 neutralizing antibodies (20 ~g/ml) to the assay. To generate standard curves of TGF-/3 isoform activity in the assay, serial dilutions of TGF-/~I, -/~2, or -/~3 (1.5-800 pg/mi) in N2 media were incubated with the transfected MLEC. After an overnight incubation at 37°C in a 5% CO2 incubator, the MLEC were washed with PBS and extracted with iysis buffer (Analytical Luminescence Laboratory, San Diego, CA). The cell lysates were analyzed for luciferase activity using lueiferin substrate (Analytical Luminescence Laboratory) and a luminometer (ML1000; Dynatech Laboratories Inc., Chantilly, VA). In parallel, we also performed a series of controls with each of the antibodies used in these assays to determine their specificity and efficiency in neutralizing each of the TGF-/~ isoforms. On average, the anti-TGF-~l,2,3 monoclonal antibody inhibited 79%, 89%, and 83% of purified TGF4~I, TGF-/Y2, and TGF-B3, respectively. The anti-'rGF-/31 antibody inhibited 80%, 3%, and 5% of purified TGF-/~I, TGF-/T2, and TGF-33, respectively; the anti-TGF-ff2 antibody inhibited 3%, 92%, and 20% of purified TGF/~1, TGF-/T2, and TGF-/~3, respectively; and the anti-'l'GF-/33 antibody inhibited 6%, 39%, and 94% of purified TGF-~I, TGF-/~2, and TGF-/33, respectively. (Thus, the anti-TGF-/~l antibody was highly specific; the antibodies to TGF-ff2 and TGF-/33 displayed some cross-reactivity with recombinant TGF-~3 and TGF-ff2, respectively). The calculated concentrations of TGF-/~ isoforms were corrected for this cross-reactivity. Effects of TGF-~I on the Expression of Schwann Cell Markers Primary rat Schwann cells were expanded in tissue culture flasks (T-75 Primaria; Falcon, Oxnard, CA) or poly-L-lysine-coated flasks in D media (DME containing 10% FBS and 2 mM glutamine) supplemented with 2/tM forskolin and 10/tg/ml crude glial growth factor prepared as described (Porter et al., 1986). Once confluent, the cultures were maintained in D media for at least 7 d before addition of growth factors. To examine the effects of TGF-/31 and forskolin on Schwann cell protein expression, flasks of Schwann cells were fed with either N2 media or D media containing one of the following: 1 ng/ml TGF-/~1, 10 ng/mi TGF-/31, 10/tM forskolin, or a combination of 10 ng/ml TGF-~I and 10/tM forskolin for 7 d. Control cultures were fed the corresponding media containing an appropriate amount of TGF-~I diluent. Cultures receiving N2 media or D media containing the added factors were fed either twice or three times, respectively, during the 7 d of treatment. In the N2 media experiments, three flasks were used for each treatment, and four flasks were maintained as controls. Only one flask was used per culture condition in the D media experiments. Expression of Schwann cell markers was determined by immunoblot analysis of cell lysates. To prepare lysates, the cultures were washed with PBS, scraped into lysis buffer (95 mM NaC1, 25 mM Tris-Cl, pH 7.4, 10 mM EDTA, 2% SDS, 1 mM PMSE and 10 mg/ml each of antipain, pepstatin A, and leupeptin), incubated in a boiling water bath for 5 rain, and then briefly sonicated. The lysates were spun in a microfuge to remove insoluble material, and the protein concentrations of the cleared supernatants were Einheber et al. TGF-{3 1 Regulates Axon/Schwann Cell Interactions 445 on A uust 4, 2017 jcb.rress.org D ow nladed fom determined using the Micro BCA method (Pierce Chemical Co., Rockford, IL). 75/~g of protein from each lysate was fractionated on a 5-15% SDS polyacrylamide gel and blotted onto nitrocellulose. Blots were incubated with primary antibodies followed by 125I-labeled protein A (Amersham Intl.) and exposed for autoradiography. Quantitation of the immunoreactive bands on the blots was performed on a Phosphorlmager (Molecular Dynamics, Inc., Sunnyvale, CA). Effects of TGF-~I on Myelination To examine the effects of TGF-~I on the differentiation of Schwann cells, cocultures of Schwann cells and DRG neurons were grown under myelinpromoting conditions (standard media containing ascorbic acid) in the presence or absence of 1 or 10 ng/mi TGF-~I. TGF-~I was added to cultures either when they were initially switched to myelin-promoting conditions or, alternatively, after they bad been maintained in myelinating conditions for 2 d. The media for control cultures and cultures treated with 1 ng/ml TGFl~l were supplemented with the TGF-~I diluent (i.e., I mg/ml low endotoxin BSA in 5 mM HC1) equal in amount to that added to the cultures treated with 10 ng/ml TGF-~I. Cultures were subsequently given the standard media containing ascorbic acid with or without TGF-~I every 2 or 3 d. After 8 d of growth under myelinating conditions, the cultures were processed for immunofluorescence or electron microscopy (described below). To determine the number of myelin segments present in the control and TGF/~l-treated cultures, the coverslips were immunostained for MBP and examined by epifluorescence on a microscope (Axiophot; Carl Zeiss, Inc., Thornwood, NY). A transparent grid was placed over the coverslips to facilitate counting the MBP-positive myelin segments on each coverslip. Statistical analyses were performed using the StatView program (Abacus Concepts, Inc., Berkeley, CA). Effects of TGF-~I on Schwann Cell Proliferation The proliferation of Schwann cells grown in the absence or presence of neurons was determined using a BrDU nuclear labeling assay. To investigate the effects of TGF-/31 on Schwann cells grown alone, 100,000 Schwann cells were plated onto poly-L-lysine-coated glass coverslips in standard media. The next day, the cultures were fed standard media with or without 1 or 10 ng/ml TGF-~I; control cultures and cultures treated with 1 ng/mi TGF-/31 were also supplemented with "rGtL01 diluent equal in amount to that added to the cultures treated with 10 ng/ml TGF-31. After •72 h, the culture media was supplemented with BrDU (10 ~tM final concentration). The cultures were incubated with BrDU in a 7% CO2 incubator at 35°C for 3.5 h. Cultures were washed in Dulbeoco's PBS, fixed in 100% methanol at -20°C for 15 rain, and incubated in 2 N HC1 for 10-30 rain. The HCI solution was removed and the cultures neutralized by two 5-rain incubations in 0.15 M sodium borate buffer, pH 8.4, and several washes with LI5 media (GIBCO BRL, Gaithersburg, MD). The cultures were then blocked with LI5 media containing 10% serum for 30 rain, and they wore incubated with fhiorescein-conjngated anti-BrDU antibody for 1 h at room temperature. Coverslips were washed in PBS and mounted on glass slides in Citifluor (Citifluor Ltd., London, U.K.) containing 1 mg/ml Hoechst dye. The effects of TGF-~I on the proliferation of Schwann cells in coculture with neurons were also investigated. In these studies, neuron cultures were seeded with 200,000 Schwann cells in standard media and maintained in N2 media for an additional 3 d. The cultures were then fed standard media containing ascorbic acid to promote myelination. TGF-/31 was added to cultunes either when they were switched to myelin-promoting media or, alternatively, after they had been maintained in this media for 2 d. The BrDU proliferation assay was performed ,,o20 h after the addition of TGF-~ to the cultures. Proliferation assays were also performed on cocultures maintained for 7 d in standard media after seeding neurons with Schwann cells. These cultures were switched to standard media containing ascorbic acid with or without TGF-/31 for 20 h before the proliferation assay was performed. BrDUand Hoechst dye-labeled nuclei in random fields of the coverslips were photographed using a Zeiss Axiophot microscope and slide film (Ektachrome 1601"; Eastman Kodak Co., Rochester, NY). At least five or six random fields were photographed, typically using a 20x objective. The number of BrDUand Hoechst dye-labeled nuclei in each field were then Counted in a blinded manner from the Ektachrome slides projected on a slide Viewer. In the case of Schwann cells grown alone, ',,1,500 cells per condition were counted; in the case of the Schwann cell/neuron cocultures, in excess of 2DO0 Cells per .condition were counted. lmmunofluorescence Microscopy Cultures were processed for immunofluorescence microscopy as.described previously (Einheber et al., 1993).