trkB encodes a functional receptor for brain-derived neurotrophic factor and neurotrophin-3 but not nerve growth factor

Cell, Vol. 65, 665-693,

May 31, 1991, Copyright

0 1991 by Cell Press

t&B Encodes a Functional Receptor for Brain-Derived Neurotrophic Factor and Neurotrophin-3 but Not Nerve Growth Factor Stephen P. Squinto, Trevor N. Stitt, Thomas H. Aldrich, Samuel Davis, Stella M. Bianco, Czeslaw Radziejewski, David J. Glass, Piotr Masiakowski, Mark E. Furth, David M. Valenzuela, Peter S. DiStefano, and George D. Yancopoulos Regeneron Pharmaceuticals, Inc. 777 Old Saw Mill River Road Tarrytown, New York 10591

Summary A variety of findings seem to functionally link brainderived neurotrophic factor (BDNF) )’and neurotrophin3 (NT-3), while distinguishing both of these factors from the third member of the neurotrophin family, nerve growth factor (NGF). Here we demonstrate that all three of these neuronal survival molecules bind similarly to the low affinity NGF receptor, but that BDNF and NT-3, unlike NGF, do not act via the high affinity NGF receptor. However, both BDNF and NT-3, but not NGF, bind to full-length and truncated forms of a receptor-like tyrosine kinase, trkB, for which no ligand had previously been identified. In addition to binding BDNF and NT-3, trkB can mediate functional responses to both of these neurotrophins when it is expressed in PC12 cells, although BDNF appears to be the more effective ligand. Thus trkBencodes an essential component of a functional receptor for BDNF and NT-3, but not for NGF. Further evidence predicts the existence of additional functional receptors for the neurotrophins. Introduction The development and maintenance of the vertebrate nervous system depends on proteins, termed neurotrophic factors, originally defined by their ability to support neuronal survival (Snider and Johnson, 1989). Neurotrophic factors have also been implicated in processes involving the proliferation and differentiation of neurons (e.g., Cattaneo and McKay, 1990; Lindsay and Harmar, 1989), and they may play additional, thus far unexplored, roles both within and outside of the nervous system. Brain-derived neurotrophic factor (BDNF) and neurotrophin3 (NT-3) have recently been molecularly cloned and shown to be structurally related to the prototypical neuronal survival molecule, nerve growth factor (NGF) (Leibrock et al., 1989; Hohn et al., 1990; Maisonpierre et al., 1990a; Rosenthal et al., 1990; Ernfors et al., 1990; Jones and Reichardt, 1990). These three proteins (designated “neurotrophins”) comprise a family, each member of which shares about 55%-80% amino acid sequence identity with the others. By contrast, the neurotrophins display no apparent structural homology with a fourth neurotrophic factor, ciliary

neurotrophic factor (CNTF) (Lin et al., 1989; Stockli et al., 1989). Understanding the biological roles of these neurotrophic factors requires characterization of the receptor and signal transduction pathways they use to exert their effects. The receptor and signal transduction pathways utilized by NGF remain unclear, although they have been extensively studied, in large part due to the availability of a pheochromocytomacell line (PC1 2) that differentiates in response to NGF (Greene and Tischler, 1978). Binding analyses with NGF have identified two classes of receptors on responsive neurons and PC12 cells, one with a K,, of 1Om8to 10eg M and the other with a K,, of lo-lo to 10-l’ M; cross-linking experiments have associated these sites with cell surface polypeptides of approximately 80 and 140 kd, respectively (Sutter et al., 1979a; Schecter and Bothwell, 1981; Massague et al., 1981; Bernd and Greene, 1984; Hosang and Shooter, 1985; Meakin and Shooter, 1991 a). A transmembrane protein (designated here as “LNGFR” for low affinity NGF receptor) with properties sufficient to account for the low affinity receptor sites has been identified and molecularly cloned (Chao et al., 1986; Radeke et al., 1987). By itself, the LNGFR does not appear to transduce a functional response to NGF, and functional responses to NGF correlate with binding to the high affinity NGF receptor sites (Zimmerman et al., 1978; Sutter et al., 1979b; Bernd and Greene, 1984). The potential role of the LNGFR in modulating binding to form a high affinity NGF binding site remains an area of active investigation (Hempstead et al., 1989; Klein et al., 1991) although it seems clear that the 140 kd receptor that can be cross-linked to NGF under high affinity conditions does not include the LNGFR (Meakin and Shooter, 1991a). Recent studies have further characterized the 140 kd component of the high affinity NGF receptor (designated “HNGFR”). This protein is phosphotylated on tyrosine in response to NGF, and apparently contains intrinsic tyrosine kinase activity (Meakin and Shooter, 1991a, 1991 b). Furthermore, early intermediates in tyrosine kinase-activated signal cascades, the ERK kinases (also known as the MAP2 kinases), are rapidly activated and phosphorylated on tyrosine in response to NGF (Boulton et al., 1991). Thus, like manyothergrowth factor responses, NGFsignal transduction may initiate by the activation of a receptorlinked tyrosine kinase. Several emerging lines of evidence now point strongly to the protein encoded by the protooncogene trk (designated here as trkA to distinguish it clearly from the related gene r&B) as a critical component of the HNGFR. The trkA gene encodes a 140 kd glycoprotein that is expressed preferentially within the nervous system; this protein has the structural features expected of a receptor tyrosine kinase (Martin-Zanca et al., 1989,199O). Although for many years no ligand had been identified for this putative receptor, recent work has revealed that the trkA protein is rapidly phosphorylated on tyrosine upon exposure of PC12 cells to NGF (Kaplan et al., 1991; Klein

Cdl 556

A.

COS ---NGF

0. BDNF

NT3

-+-+-+

COS + LNGFR ---NGF

BDNF

NT3

-+-+-+

C. CDS+LNGFR vs. A875

cos+ .---

LNGFR A875 -+ -+

200 -

12

34

56

123456

Figure 1. All Three Neurotrophins LNGFR Expressed in COS Cells

‘12 Specifically

Cross-Link

34 to the

(A) None of the neurotrophins display competable cross-linking to COS cells transfected with control vector, pCMX. The radiolabeled ligand utilized for each pair of lanes is indicated at top of lanes (each radiolabeled ligand is estimated to be at a concentration between 0.1 to 0.25 nM); “-” indicates absence and “+” indicates presence of unlabeled homologous ligand at a concentration of 500 nM. Notable bands seen with radiiabaled NGF varied from experiment to experiment and were not competable with unlabeled NGF, as indicated. (B) All three radiolabeled neurotrophins display competable crosslinking (resulting in a complex of approximately 100 kd as expected for LNGFR) in COS cells transfected with pCMX-LNGFR; lanes marked as in (A). (C) Cross-linked species in COS cells transfected with pCMX-LNGFR comigrates with cross-linked species in human A575 melanoma cells. Radiolabeled ligand used in this panel was BDNF; cells used for crosslinking are indicated at top of each pair of lanes, and “-” and ‘+” are as in (A).

et al., 1991) and that trkA directly binds to NGF when expressed in heterologous ceils, even in the absence of the LNGFR (Klein et al., 1991). In contrast to the extensive studies with NGF, the receptors and signal transduction pathways utilized by the related neurotrophins have only recently begun to be explored. BDNFappears to bind to the LNGFR with an aff inky similar to that of NGF (Rodriguez-Tebar et al., 1990). However, the findings that BDNF and NGF can act on different neurons and that NGF-responsive neurons do not generally express high affinity BDNF receptors suggest that BDNF utilizes a different high affinity receptor than NGF (Rodriguez-Tebar and Barde, 1988). NT-3 receptors have not yet been described. A variety of findings seem to link BDNF and NT-3, while distinguishing both of these neurotrophins from NGF. Although the three neurotrophins display very different patterns of expression in the periphery, striking reciprocal relationships were observed between the expression of BDNF and NT-3, but not NGF, during development of the central nervous system; NT-3 is expressed more prominently early and BDNF more prominently late during the maturation of some of the same brain regions (MaisonPierre et al., 1990b). Interestingly, the distribution profiles of BDNF and NT-3 (but not NGF) in various adult brain regions are quite similar (Maisonpierre et al., 199Ob). Both BDNF and NT-3 elicit prominent neurite outgrowth from explanted dorsal root and nodose ganglia, whereas NGF

induces strong outgrowth from dorsal root and sympathetic, but not nodose, ganglia (Maisonpierre et al., 199Oa). These findings led to the suggestion that BDNF and NT-3 might, in some cases, act on the same neuronal populations (Maisonpierre et al., 1990b). Furthermore, the finding that BDNF and NT-3 (but not NGF) are the most highly conserved growth factors yet described led to the suggestion that both of these factors might interact with multiple receptors (Maisonpierre et al., 1991). The trkB gene encodes a transmembrane growth factor receptor-like protein closely related to trkA (Klein et al., 1989, 1990; Middlemas et al., 1991); these receptor-like proteins are most similar within their tyrosine kinase domains, although there is substantial homology within their extracellular domains. A variety of differentially expressed f&B transcripts exist, some of which encode a truncated form of this protein that lacks the intracytoplasmic tyrosine kinase domain. As with WA, the expression of t&B is also largely restricted to the nervous system, although t&B expression appears to be more widespread within both the peripheral and central nervous systems. Interestingly, low level t&B expression outside of the nervous system (most notable in lung and muscle) parallels low level BDNF expression seen outside of the nervous system (Klein et al., 1989; Maisonpierre et al., 199Ob). Here we demonstrate that all three neurotrophins bind similarly to the LNGFR, but that BDNF and NT-3, unlike NGF, do not act via the HNGFR. In contrast, we find that both BDNF and NT-3 bind to the t&B gene product, whereas NGF does not bind detectably to this protein. In addition, trkB mediates functional responses to both BDNF and NT-3 when it is transfected into PC12 cells, although BDNF appears to be the more effective ligand in these assays. Finally, the identification of a human neuroblastoma cell line that responds to BDNF but not NT-3, and lacks tr/rB expression, predicts the existence of additional functional receptors for the neurotrophins. Results All Three Neurotrophins Bind to the LNGFR, but BDNF and NT-3 Do Not Act via the HNGFR To determine whether NT-3, like NGF and BDNF (Rodriguez-Tebar et al., 1990) could bind to the LNGFR, we examined all three neurotrophins for their ability to be chemically cross-linked to the LNGFR protein expressed transiently in COS cells. Each of the three radiolabeled neurotrophins could be specifically cross-linked to a species of the molecular weight expected for the LNGFR (Figure 1B; cross-linked complex previously reported to be approximately 100 kd by Hosang and Shooter, 1985). The cross-linked product was not observed on control COS cells that were not expressing the LNGFR protein (Figure 1A). Furthermore, as expected for specific binding, the appearance of the cross-linked products could be competed effectively by an excess of the corresponding unlabeled neurotrophin (Figure 1 B). Competable cross-linking to a polypeptide of the correct size was also observed in

trkB Is a Functional 887

Receptor

for BDNF

and NT-3

BSA Figure 2. Induction of Neurite Outgrowth and Immediate-Early Gene Expression in PC12 cells in Response to NGF but Not BDNF or NT-3

,fos (2.2

Jun

t

(2.7

kb)

kb)

._..

a cell line (A875 melanoma) known to stably express large amounts of the LNGFR (Figure 1C). Having established that all three neurotrophins bind to the LNGFR protein, we assessed the ability of BDNF and NT-3 to function via low or high affinity NGF receptors. The PC12 cell line, which expresses both classes of NGF receptors, displays prominent responses including both neurite outgrowth and the transcriptional induction of a set of “immediate-early” genes in response to NGF (Greene and Tischler, 1978; Greenberg et al., 1985; see Figures 2A and 28). However, when PC12 cells were incubated with BDNF or NT-3 at concentra0ons either below or exceeding that known to be saturating for responses to NGF, we observed neither morphological changes (under conditions optimal for neurite outgrowth, as described below) nor the induction of immediate-early gene expression (Figures 2A and 28). This implies that neither low nor high affinity NGF receptors suffice for functional responses to BDNF and NT-3, and that such responses require at least one receptor component not normally found on PC1 2 cells. BDNF and NT-3, but Not NGF, Bind to Full-Length and Truncated Forms of trkB trkB, a 145 kd glycoprotein that is a potentiaPcomponent of the receptor for BDNF and/or NT-3, is not expressed in PC12 cells (data not shown; Kaplan et al., 1991); cell lines expressing trkB have not been described. A full-length rat trkB cDNA was isolated and transiently expressed in COS cells, on which cross-linking experiments were carried out with radioiodinated neurotrophins. As shown in Figure 3A, both BDNF and NT-3, but not NGF, could be cross-linked

(A) PC12 cells were cultured as recommended by Greene et al. (1987) in the presence of 100 @ml BSA and varying concentrations (l-200 ng) of NT-3, BDNF, or NGF; results obtained with 100 nglml of each neurotrophin are depicted. (B) fos and jun transcripts (2.2 and 2.7 kb, respectively) were identified by Northern analysis of total cellular RNA prepared from PC1 2 cells cultured as above in the absence of exogenous neurotrophic factors, and then treated for 30 min with 100 nglml BSA, NT-3, BDNF, or NGF, as indicated.

to a polypeptide of approximately the size expected for the frkB gene product; some heterogeneity in the size of this cross-linked species was observed, as has been previously reported for the HNGFR cross-linked to NGF (Meakin and Shooter, 1991a). Cross-linking of labeled BDNF or NT-3 to the presumptive t&B gene product was not observed in the presence of excess unlabeled homologous ligand (Figure 3A, lanes labeled “+“). To verify that the protein cross-linked to BDNF or NT-3 actually corresponds to the trkB gene product, and to determine whether certain trunctated forms of the protein are capable of binding ligands, a tfkB deletion mutant lacking the intracytoplasmic protein-tyrosine kinase domain was constructed and expressed in COS cells. As illustrated in Figure 38, radiolabeled ligand (in this case NT-3) was cross-linked efficiently to a surface component of cells expressing the truncated trkB product; furthermore, the marked shift in mobility (corresponding to about 35 kd) observed between the cross-linked species obtained with full-length versus truncated trkB proteins agreed well with the known size of the deletion. BDNF and NT-3, but Not NGF, Compete for Binding to trkB The specificity of binding of the neurotrophins to the LNGFR and to trkB was compared in competition assays. Each of the three unlabeled neurotrophins specifically blocked the cross-linking of radiolabeled NT-3 to the LNGFR when present at 500 nM but not at 1 to 5 nM (Figure 4A). Consistent with these results, direct binding inhibition

Cdl 888

A.

trkB NGF -w-

9. trkB(del) BDNF

1234 Figure 3. BDNF and NT-3, Expressed in COS Ceils

NT3

56 but Not NGF,

NT3

1 2 Bind Specifically

to trkB

(A) COS cells were transfected with pCMX-trkB and chemically crosslinked to each of the three radiolabeled neurotrophins. The radiolabeled ligand utilized for each pair of lanes is indicated at top of lanes (each radiolabeled ligand is estimated to be at a concentration between 0.1 and 0.25 nM); “-” indicates absence and “+” indicates presence of unlabeled homologous ligand at a concentration of 500 nM. Brace and asterisk indicate cross-linked species corresponding to trkB protein (approximately 180 to 180 kd). (B) Cross-linking of radiolabeled NT-3, in the absence (“-3 or presence (“+“) of unlabeled NT-3 at 800 nM, to COS cells transfected with an expression vector for a truncated form (lacking the tyrosine kinase domain) of trkB (pCMX-trkB(del)). Brace and asterisk indicate crosslinked species corresponding to truncated version of trkB (120 to 150 kd, as expected for this deletion mutant).

studies revealed that the binding of radiolabeled NT-3 to the LNGFF? expressed on COS ceils was competed similarly by each of the three unlabeled neurotrophins (Figure 4C); the competition curves suggest dissociation constants in the nanomolar range for all three neurotrophins, extending previous observations for NGF and BDNF (Rodriguez-T6bar et al., 1990). In contrast to the rather high levels of unlabeled ligands required to specifically block cross-linking to the LNGFR, lower levels of unlabeled BDNF and NT-3 effectively prevented binding of radiolabeled NT-3 to trkB (Figure 48). Direct binding inhibition analysis verified that both BDNF and NT-3 compete for binding to trkB, with BDNF the more effective competitor (Figure 4D). NGF, even at 500- to IOOO-fold molar excess, did not compete for binding of radiolabeled NT-3 to trkB in either assay (Figures 48 and 4D). Because the amounts of unlabeled BDNF and NT-3 required to inhibit the binding of radiolabeled NT-3 to trkB were lower than those necessary to block binding of NT-3 to the LNGFR, our data suggest that trkB displays somewhat higher affinity for both BDNF and NT-3 than does the LNGFR. trk6 Mediates to Both BDNF When exposed morphological

Neurite Outgrowth In Response and NT-3 in PC12 Cells to NGF, PC1 2 cells display a characteristic response, neurite extension, indicative of

differentiation to a more mature neuronal phenotype. As demonstrated above, these cells do not respond to either BDNF or NT-3. To test whether trkB could mediate a biologically relevant response to BDNF or NT-3, PC1 2 cells were transiently transfected with a MB expression vector (pCMX-trkB) and incubated with each of the neurotrophins. To minimize background, the transfected cells were cultured on standard tissue culture plastic rather than either collagencoated or Primaria surfaces; under these conditions, which are suboptimal for neurite extension (Greene et al., 1967; Chen et al., 1990), NGF induced rather short neurites from control PC12 cells (not shown) as well as from the MB-transfected cells (Figure 58). No cells with neurites were seen in the MB-Vansfected cultures in the absence of added neurotrophic factor (Figure 5C). However, many cells in the tr/rB-transfected PC12 cultures displayed robust neuritic outgrowth in response to BDNF, while fewer but significant numbers of cell8 differentiated in response to NT-3 (Figure8 5D and 5E). No cell8 with neurites were seen in PC12 cells transiently transfected with control vectors and treated with BDNF or NT-3 (data not shown). As a positive control to assess transfection efficiency, the PC12 cells were transfected with an activated H-ras gene, which has been shown to induce ligand-independent differentiation’of these cells; the number of differentiated cells seen in the ras-transfected cultures indicates the number of transiently transfected PC12 cells in the cultures (L. A. Greene, personal communication). Since the number of ,differentiated cells observed following BDNF treatment of pCMX-trkB-transfected PC12 cells is comparable to the number of differentiated cells found in the ras-transfected populations (Figures 5A and 5E; number of differentiated cells obtained with each treatment is indicated [see legend]), our data suggest that every PC1 2 cell expressing trkB can respond to BDNF. However, a smaller subset of trkssxpressing PC12 cells responded to NT-3 as measured by neurite outgrowth (Figure 5D); dOSf?-re8pOtISe studies suggest that BDNF is at least 10 times more effective than NT-3 in these assays (not shown). As depicted in Figure 5, it was striking that the extensive neuritic outgrowth Seen in ras-transfected PC1 2 cells or in pCMX-trkB-transfected PC1 2 cells subjected to BDNF or NT-3 was qualitatively different than the blunted neuritic outgrowth normally seen in response to NGF under these culture conditions. SH-SYSY Human Neuroblastoma Cells Respond to BDNF, but Not NT-3, and Do Not Express trkB: Evidence for Another Neurotrophin Receptor? In contrast to the large number of NGF-responsive cell lines described, no cell lines responsive to BDNF or NT-3 have been identified. We have used the induction of immediate-early gene expression as an assay (Squint0 et al., 1990) to search for neuronal tumor cell lines responsive to BDNF, NT-3, and NGF. One human neuroblastoma cell line (SH-SY5Y) was found to express c-fos mRNA in response to both NGF (as previously described) and BDNF, but not NT-3 (Figure 6A). Further studies have verified that BDNF, but not NT-3, has additional functional effects on

trkB Is a Functional

Receptor

for BDNF

and NT-3

009

A. LNGFR binding

to NT-3

B.

trkB binding cold

rlnln

::

,123456789

123456789 LNGFR

NT3

BDNF oF-n::

NGF 0 -w::

C.

to NT-3

wmoetltor

to NT-3

binding

D.

trkB

binding

(A) COS cells were transfected with pCMXLNGFR and cross-linked to radiolabeled NT-3 (estimated concentration between 0.1 and 0.25 nM); the unlabeled neurotrophin used as unlabeled competitor is indicated at the top of each triplet of lanes, with the concentration used per lane (in nM) indicated. (B) COS cells were transfected with pCMX-trkB and cross-linked to radiolabeled NT-3; ligand concentrations and lane markings are as in (A). (C and D) COS cells transfected with pCMXLNGFR (C) or pCMX-trkB (D) were used in solution binding experiments with radiolabeled NT-3 (at a concentration estimated to be between 0.1 and 0.25 nM) competed with varying concentrations of unlabeled NT-3, BDNF, and NGF, as indicated. Data represent the percentage of total cpm bound and are the average of duplicate assays.

to NT-3

-ElDW -b&Y

260

Figure 4. Cross-Linking and Binding of 9NT-3 to LNGFR Is Specifically Blocked by All Three Neurotrophins, While Cross-Linking and Binding of NT-3 to trkB Are Specifically Blocked Only by BDNF and NT-3

NT3

z m E I H z

[COMPETING

NEUROTROPHIN] 0-M

60

40 20

[COMPETING

NEUROTROPHIN] 0-Y

SHSY5Y; for example, BDNF protects these cells from certain neurotoxins (M. 6. Spina and S. P. S., unpublished data), as has been previously demonstrated for dopaminergic neurons in cultures (Hyman et al., 1991). Although this cell line has been shown to express low levels of both high and low affinity NGF receptors (Chen et al., 1990) and detectable levels of trkA mRNA (data not shown), it did not express detectable levels of trks mRNA (Figure 6B). The apparent lack of trkB expression in a BDNF-responsive ceil line, together with the failure of this line to respond to NT-3, leads us to predict either that there are additional modulators of the LNGFR or trkA which can confer BDNF responsiveness, or that yet another functional neurotrophin receptor exists which has discrete specificity for BDNF. DiscuBsion We conclude that trkB encodes an essential component of a functional receptor for BDNF and NT-3, but not for the third neurotrophin family member, NGF. Recent reports suggest that the trkA proto-oncogene, the closest known

relative of trk8, encodes a component of the HNGFR (Kaplanetal., 1991; Kieinetal., 1991; Meakinetal., 1991a, 1991b). Our observations that normal PC12 cells do not respond to BDNF or NT-3 imply that trkA, which is expressed in PC12 ceils, is uniquely activated by only one known member of the neurotrophin family, NGF. We find that BDNF and NT-3 bind to trkB in the absence of the LNGFR, and our data suggest that this binding is of higher affinity than the binding of these ligands to the LNGFR. Similarly, Klein et al. (1991) reported that NGF can bind to trkA with relatively high affinity in cells that do not express the LNGFR. We extend previous findings (Rodriguez-Tebar et al., 1990) by demonstrating that all three neurotrophins bind to the LNGFR with approximately equal, albeit relatively low, affinity; the conservation of this property suggests a significant biological role. The LNGFR may modulate the binding of each of the neurotrophins to its appropriate trk receptor, as has been suggested by previous studies (e.g., Hempstead et al., 1969). Alternatively, the LNGFR may mediate signal transduction via an independent pathway, or it may not be directly involved in initiating signal transduction. For example, it may act to

b

zr/rB+MGF (ND)

Figure 5. PC12 Cells Transfected with pCMX-trkB Differentiate in the Presence of BDNF and NT-3 (A) PC12 ceils were transiently transfected with control plasmid pT24ras and treated with 100 rig/ml BSA to determine the number of transiently transfected cells in each experiment (see text for details). (B-E) PC12 cells were transiently transfected with pCMX-trkB and treated with 100 ng/ ml NGF (B), BSA (C), NT-3(0), or BDNF (E). Numbers (in parentheses) at the bottom of each panel indicate the number of differentiated PC12 cells (i.e., cells having neurites more than twice the length of the cell body) observed per 35 mm wefl following each treatment. Although absolute numbers varied in the three transfections performed, the ratios of differentiated cells obsenred foffowing the different treatments remained constant for three independent experiments; the percentage of differentiated cells seen in the ras-transfected populations in the three independent experiments ranged from 3%~5%. No dffferentiated cells were observed in pCMX-trkB-transfected PC12 ceffs treated with BSA (three seperate experiments). After electroporation the cefls were plated directly on plastic without precoating to eliminate any background neurite outgrowth, as described in the text and in the Experimental Procedures.

localize, concentrate, or trap the neurotrophins on the surface of LNGFR-expressing cells. In this regard it is of interest that neuronal supporting cells (such as Schwann cells) that do not respond to NGF express the LNGFR and up regulate it in response to injury (Johnson et al., 1988), perhaps providing a fixed matrix or path for the concentration and presentation of neurotrophins to regenerating neurons. Alternatively, the LNGFR may act as a “clearance” receptor that reduces free or circulating levels of the neurotrophins; the LNGFR is widely distributed both in the nervous system and in peripheral tissues (Bothwell, 1991), and secreted forms of the LNGFR (DiStefano and John-

son, 1988) may aid in such clearance mechanisms. In regard to nonsignaling roles for the LNGFR, related mechanisms specific for BDNF and NT-3 can be proposed based on the presence of truncated forms of trk6. The colocalization of BDNF and truncated MB transcripts outside of the nervous system (most prominently in lung and skeletal muscle) raises intriguing questions concerning the role of BDNF, trkB, and other potential BDNF receptors in nonneural tissues. The fact that BDNF and NT-3 can act on the same recep tor, trkB, is consistent with our previous suggestions that the distributions and overlapping neuronal specificities of

;ry

Is a Functional

Receptor

for BDNF

and NT-3

-fos (2.2

kb)

- 285

trk B - 18s

Figure 6. The Human Neuroblastoma SH-SY5Y and BDNF but Not NT-3, and Does Not Express

kesponds trkB

to NGF

(A) fos transcripts were identified by Northern analysis of total cellular RNA prepared from SHSYSY cells cultured in the absence of added factor, and then treated for 30 min with 199 @ml BSA, NGF, NT-3, or BDNF as indicated. (B) t&B transcripts can be detected by Northern analysis using 10 ug of total cellular RNA from adult rat cerebellum (lane CB), but are not detectable in 10 ug of total cellular RNA from neuroblastoma SH-SYSY. The complete coding region of f&B was used as a probe, which identifies multiple f&B transcripts (Klein et al., 1989); the cerebellum lane appears as a smear because of gross overexposure in the attempt to detect transcripts in SH-SYSY.

BDNF and NT-3 particularly link the roles of these two neurotrophins, at least in the central nervous system. During the maturation of different brain regions marked by decreasing NT-3 level8 and increasing BDNF levels, the expression of trkB remain8 relatively constant (S. P. S., P. M., and G. D. Y., unpublished data). However, the finding that BDNF is a more effective ligand for trkB than is NT-3, at least in one type of functional assay, together with the identification of a cell line that respond8 to BDNF but not NT-3 and does not detectably express tf/rB, strongly suggests that the89 two neurotrophins are not entirely interchangeable. These data predict the existence of additional receptors, or modulatory components, that allow for distinct responses to either BDNF or NT-3. Such modulation may explain, for example, the differing effects of NT-3 and BDNF on sympathetic neurons (Maisonpierre et al., 199Oa). Sympathetic neuron8 are known to express t&B, although the available in situ hybridization data do not distinguish between the presence of functional or nonfunctional frke transcripts in these neurons (Klein et al., 1989). An evolutionary comparison of BDNF and NT-3 led us to predict that these two neurotrophins were strictly conserved to maintain their specific interactions with multiple receptors (Maisonpierre et al., 1991). Although NGF is well conserved evolutionarily compared with most secreted factors, it does not display the striking conservation that BDNF and NT-3 do, perhaps suggesting that it does not interact with as many receptors as the other two known neurotrophins. Our findings place the neurotrophins in the

growing class of receptor-ligand systems in which multiple receptors each bind to several different related ligands (reviewed in Cross and Dexter, 1991). The binding and activation of receptor tyrosine kinases by the neurotrophins reveal that these factors utilize signaling pathways very similar to those activated by mitogenie growth factors. This finding is consistent with recent data that neurotrophic factors can act a8 mitogens in certain contexts (e.g., Cattaneo and McKay, 1990) but also indicates that signals that initiate with the activation of receptor tyrosine kinases normally integrate into nonmitogenie transduction pathways in neurons. Despite differences in ultimate sequelae (i.e., mitogenesis vs. survival or differentiation), at least 8ome of the early intermediates in tyrosine kinase signaling cascades, such a8 the ERK family of protein kinases, are similarly activated in both neuronal and nonneuronal cells (Boulton et al., 1991). Other examples in which activation of receptor-like tyrosine kinases in neuronal cells leads to nonmitogenic sequelae include the Drosophila sevenless gene product, where activation via a nondiffusible ligand is required for the differentiation of photoreceptor cells (Basler et al., 1991). It is becoming increasingly clear that the signal transduction mechanisms utilized by neurons fundamentally resemble those employed by other cell types. Studies of neuropeptides and neurotransmitters have focused attention on the role of G protein-coupled receptors in modulating neuronal activity. The studies of neurotrophin receptors now implicate receptor tyrosine kinases in the action of neuronal survival molecules. We have found that a neurotrophic factor unrelated to the neurotrophins, CNTF, utilizes yet another type of receptor system (Davis et al., submitted). A key goal for the future will be to understand how these diverse signal transduction mechanisms are interpreted and integrated in the context of a neuron. Experimental

Procedures

Cell Culture, Neurotrophins, and lodination of Neurotrophlns COS-MI cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS), 1% each of penicillin and streptomycin (P/S), and 2 mM glutamine in an atmosphere of 5% Con. PC12 cells were cultured in DMEM with 6% FBS and 8% horse serum, P/S, and 2 mM glutamine on Costar tissue culture plates in an atmosphere of 7.5% CO,; PC12 cells obtained from Dr. L. A. Greene’s laboratory were utilized in experiments depicted in Figure 2, and PC12 cells obtained from Dr. E. M. Shooter’s laboratory were utilized in experiments depicted in Figure 5. The human neuroblastoma cell line SH-SY5Y was cultured in Eagle’s minimal essential medium with 10% FBS, P/S, and 2 mM glutamine. Murine 2.5s NGF was obtained from Bioproducts for Science (Indianapolis, IN). Both human BDNF and NT-3 were produced in CHO cells and purified from CHO cell conditioned media to homogeneity as assessed by silver-stained polyacrylamide gels and amino acid sequence analysis. Purified neurotrophins (NGF, BDNF, and NT-3) were all iodinated using the lactoperoxidase method as previously described(Hempsteadet al., 1989). lodinated neurotrophins weresep arated from unincorporated ‘PI by using Centriflo CF58A filters (Amicon, Beverly, MA) and gel filtration (5299) column chromatography. Cloning Rat t&B, Mammalian Expresaion Constructs, and Transient Tranefections A full-length rat tr/rB cDNA clone was obtained by screening a rat brain cDNA library in the kZAP2 vector (Stratagene) with rat t&B-specific

Cdl 892

oligonucleotides corresponding to the most 5’and 3’coding regions of tdd3.Both the human LNGFR (Johnson et al.. 1988) and rat frkB cDNAs were subcloned into the mammalian expression vector pCMX (Davis et al., submitted) to generate pCMX-LNGFR or pCMX-trkB, respectively. pCMX-trkB(del) was generated by digesting the pCMX-trkB plasmid with Apal (which cuts just after the f&B transmembrane domain) and Notl (which cuts just after the t&B coding region in vector sequences), blunting these ends, and religating the plasmid; the t&B coding region generated includes all of the extracellular and transmembrane domains of trke but lacks the C-terminal 320 amino acids. COSM5 cells were transiently transfected with either the pCMXLNGFR, pCMX-trkB, or control vector (pCMX) by the DEAE-dextran transfection protocol. Briefly, COS-MB cells were plated at a density of 1.5 x 1Q cells per 100 mm plate 24 hr prior to transfection. For transfechon, the cells were cultured in serum-free DMEM containing 400 ug/ml DEAE-dextran, 1 uM chloroquine, 2 mM glutamine, 20 ug/ ml insulin, 5 us/ml transferrin, 33 nM sodium selenite, and 5 ug of the appropriate DNA for 3 hr and 15 min at 37OC in an atmosphere of 5% CO,. The transfection medium was aspirated and replaced with phosphate-buffered saline with 10% DMBO for 2 min. Following this DMSO “shock,” the COS-MS cells were placed into DMEM with 10% FBS, P/S, and 2 mM glutamine for 48 hr. PC12 cells were transiently transfected by electroporation. Briefly, the cells were rinsed prior to transfection in ice-cold Dulbecco’s phosphate-buffered saline (calciumand magnesium-free) and then resuspended in the same buffer at a density of 1.5 x IO7 cells per ml containing 40 ug of the appropriate DNA. PC12 cells were transfected with either pCMX, pCMX-trkB, or the pT24-ras plasmid (Yancopoulos et al., 1985). The cell mixture was incubated on ice for 10 min and then quickly brought to room temperature and electroporated in a total volume of 1 ml at 1150 V/cm and 500 pF. Electroporated cells were incubated on ice for 30 min prior to plating in DMEM with 8% FBS and 8% horse serum with P/S and 2 mM glutamine; cells were plated on Costar plastic in the absence of any precoating. Forty-eight hours after transfection, the cells were treated with 100 @ml neurotrophin or BSA (see legend of Figure 5). and neurite outgrowth was scored 72 hr later. Chemical Cross-Linking Cells were harvested in phosphate-buffered saline containing 1 mM EDTA, 1 mg/ml glucose, and 25 mM HEPES (PBS-versene) and resuspended at an appropriate density (generally 1 x 10s cells per ml) in ice-cold binding buffer A (PBS containing 1 mg/ml each BSA and glucose). COS-MB cells (4 x 1 v) transfected with pCMX-trkB or pCMX vector or COSM5 cells (3 x lv) transfected with pCMX-LNGFR were incubated on ice with 1251-labeled neurotrophins (final concentration estimated to be between 0.1 and 0.25 nM) for 90 min in the absence or presence of unlabeled NGF, BDNF, or NT-3 (see Figures 1 and 3). The chemical cross-linker DSS (Pierce, Rockford, IL) was used following conditions previously described (Meakin and Shooter, 199la). The cross-linking reaction was terminated after 90 min and quenched with 12 ml of 50 mM Tris buffer containing 180 mM NaCI. Cells were centrifuged at 300 x g for 5 min and then washed twice with 12 ml of buffer A. Pelleted cells were solubilized in SDS containing 2% 2-mercaptoethanol and boiled for 5 min; radiolabeled cross-linked proteins were resolved on 7% polyacrylamkfe gels and visualized by autoradiography after exposure of the dried gel to Kodak X-Cmat film at -7OOC.

‘*%NT-3

Competltion

Binding

Assays

Binding of ‘25l-NT-3 to COSM5 cells transfected with either pCMXLNGFR, pCMX-trkB, or control vector (pCMX) was assessed on cells in suspension. Cells were harvested in PBS-versene and then resuspended in binding buffer A as described above for chemical crosslinking. Cells were incubated with ?NT-3 (estimated concentration between 0.1 and 0.25 nM) in the absence or presence of increasing concentrations of unlabeled NT-3, BDNF. or NGF ranging from 0.3 to 30 nM for NGF and between 1 and 100 nM for BDNF and NT-3 (see Figures 4C and 4D). The binding reactions were carried out for 90 min on ice. Free ‘=I was separated from bound ‘*? by quickly centrifuging (30 s spin) the reaction mixture through a sucrose gradient formed in a long-tip microcentrifuge tube and then immediately freezing the tube in a dry ice-ethanol bath. The bottom of the reaction tube was cut and then counted in a gamma counter.

RNA Isolation

and Northern

Blotting

Analysis

Total cellular RNA isolated from SH-SYSY and treated or untreated PC1 2 cells was fractionated on 1% formaldehyde-agarose gels, transferred to nylon membranes, and hybridized to a “P-labeled v-fos probe or a =P-labeled c-iun probe as previously described (Squint0 et al., 1990); probings for trkBexpression were performed using a “P-labeled rat m probe spanning a region encoding the intracytoplasmic tyrosine kinase domain.

Acknowledgments The authors would like to thank Dr. Leonard Schleifer and the entire Regeneron community for continued support and provocative scientific discussions. We would also like to thank Drs. E. M. Shooter, S. 0. Meakin, and M. V. Chao for helpful suggestions concerning iodination procedures and cross-linking techniques, Dr. U. Suter for suggestions on culturing PC12 cells on nonwated plastic, and Dr. L. A. Greene and his colleagues for helpful suggestions on ideal culture and electroporation conditions for the PC12 cells, and for suggesting the ras transfection controls. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC Section 1734 solely to indicate this fact. Received

April 17, 1991; revised

April 28, 1991.

References Basler, K., Christen, B., and Hafen, E. (1991). Ligand-independent activation of the sevenless receptor tyrosine kinase changes the fate of cells in the developing Drosophila eye. Cell 84, 1089-1081. Bernd, P., and Greene, L. A. (1984). Association of ‘2sI-nerve growth factor with PC12 pheochromocytoma cells. Evidence for internalization via high affinity receptors only and for long-term regulation by nerve growth factor of both high- and low-affinity receptors. J. Biol. Chem. 259, 15509-15518. Bothwell, M. (1991). Tissue nerve growth factor receptors. 70.

localization of nerve Curr. Top. Microbial.

growth factor and Immunol. 185,83-

Boulton, T. G., Nye, S. H., Robbins, D. J., Ip, N. Y., Radziejewska, E., Morgenbesser, S. D., DePinho, R. A., Panayotatos, N., Cobb, M. H., and Yanwpoulos, G. D. (1991). ERKs: a family of protein-serine/threonine kinases that are activated and tyrosine phosphorylated in response to insulin and NGF. Cell 85, 883-875. Cattaneo, E., and McKay, R. (1990). Proliferation of neuronal stem cells regulated by nerve growth 782-785.

and differentiation factor. Nature 347,

Chao, M. V., Bothwell, M. A., Ross, A. H., Koprowski, H., Lanahan, A. A., Buck, C. R., and Sehgal, A. (1988). Gene transfer and molecular cloning of the human NGF receptor. Science 232, 518-521. Chen, J., Chattopadhyay, B., Venkatakrishnan, G., and Ross, A. H. (1990). Nerve growth factor-induced differentiation of human neuroblastoma and neuroepithelioma cell lines. Cell Growth Diff. 7, 79-85. Cross, M., and Dexter, T. M. (1991). Growth factors in development, transformation, and tumorigenesis. Cell 84, 271-280. DiStefano, P., and Johnson, E. M., Jr. (1988). cated form of the newe growth factor receptor. USA 85,270-274.

Identification of a trunProc. Natl. Acad. Sci.

Ernfors, P., Ibanez, C. F., Ebendal. T., Olson, L., and Persson, H. (1990). Molecular cloning and neurotrophic activities of a protein with structural similarities to nerve growth factor. Proc. Natl. Acad. Sci. USA 87,5454-5458. Greenberg, M. E., Greene, L. A., and Ziff, E. B. (1985). Nerve growth factor and epidermal growth factor induce rapid transient changes in proto-oncogene transcription in PCl2cells. J. Biol. Chem. 280,1410114110. Greene, L. A., and Tischler, A. S. (1978). Establishment of a noradrenergic clonal line of rat adrenal pheochromocytoma cells which respond to nerve growth factor. Proc. Natl. Acad. Sci. USA 73, 2424-2428.

;ri

Is a Functional

Receptor

for BDNF

and NT-3

Greene, L. A., Aletta, J. M., Rukenstein, A., and Green, S.H. (1967). PC12 pheochromocytoma cells: culture, nerve growth factor treatment, and experimental exploitation. Meth. Enzymol. 747, 207-216.

MassaguB, J., Guilette, B. J., Czech, M. P., Morgan, C. J., and Bradshaw, R. A. (1961). Identification of a nerve growth factor receptor protein in sympathetic ganglia membranes by affinity labeling. J. Biol. Chem. 256, 9419-9424.

Hempstead, E. L., Schleifer, L. S., and Chao, M. V. (1969). Expression of functional nerve growth factor receptors after gene transfer. Science 243, 373-375.

Meakin, S. O., and Shooter, E. M. (1991a). Molecular investigations on the high affinity nerve growth factor receptor. Neuron 6, 153-163.

Hohn, A., Leibrock, J., Bailey, K., and Barde, Y.-A. (1990). Identification and characterization of a novel member of the nerve growth factor/ brain derived neurotrophic factor family. Nature 344, 339-341.

Meakin, S. O., and Shooter, E. M. (1991 b). Tyrosine kinase activity coupled to the high affinity nerve growth factor receptor. Proc. Natl. Acad. Sci. USA, in press.

Hosang, M., and Shooter, E. M. (1965). Molecular characteristics of nerve growth factor receptors on PC12 cells. J. Biol. Chem. 260,655 662.

Middlemas, D. S., Lindberg, R. A., and Hunter, T. (1991). WkE, a neural receptor protein-tyrosine kinase: evidence for a full-length and two truncated receptors. Mol. Cell. Biol. 77, 143-153.

Hyman, C., Hofer, M., Barde, Y.-A., Juhasz, M., Yancopoulos, G. D., Squinto, S. P., and Lindsay, R. M. (1991). BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350, 230-232.

Radeke, M. J., Misko, T. P., Hsu, C., Herzenberg, L. A., and Shooter, E. M. (1967). Gene transfer and molecular cloning of the rat nerve growth factor receptor. Nature 325, 593-597.

Johnson, D., Lanahan, A., Buck, C. R., Sehgal, A., Morgan, C., Mercer, E., Bothwell, M., andChao, M. (1966). Expression and structureof the human NGF receptor. Cell 47, 545-554. Johnson, E. M., Jr., Taniuchi, M., and DiStefano, sion and possible function of nerve growth Schwann cells. Trends Neurosci. 11, 299304.

p. S. (1966). Expres‘factor receptors on

Jones, K. R., and Reichardt, L. F. (1990). Molecular cloning gene that is a member of the nerve growth factor family. Acad. Sci. USA 87,6060-6064.

of a human Proc. Natl.

Rodriguez-T&bar, A., and Barde, Y.-A. (1966). Binding characteristics of brain-derived neurotrophic factor to its receptors on neurons from the chick embryo. J. Neurosci. 8, 3337-3342. Rodriguez-T&bar, A., Dechant, G., and Barde, Y.-A. (1990). Binding of brain-derived neurotrophic factor to nerve growth factor receptor. Neuron 4, 407-492. Rosenthal, A., Goeddel, D. V., Nguyen, T., Lewis, M., Shih, A., Laramee, G. R., Nikolics, K., and Winslow, J. W. (1990). Primary structure and biological activity of a novel human neurotrophic factor. Neuron 4, 767-773.

Kaplan, D. R., Martin-Zanca, D., and Parada, L. F. (1991). Tyrosine phosphorylation and tyrosine kinase activity of the trk proto-oncogene product induced by NGF. Nature 350, 156-160.

Schechter, A. L., and Bothwell, M. A. (1961). Nerve growth factor receptors on PC12 cells: evidence for two receptor classes with differing cyioskeletal association. Cell 24, 667-674.

Klein, R., Parada. L. F., Coulier, F., and Barbacid, M. (1969). trkB, a novel tyrosine protein kinase receptor expressed during mouse neural development. EMBO J. 8, 3701-3709.

Snider, W. D., and Johnson, Ann. Neurol. 26, 469-506.

Klein, R., Conway, D., Parada, L. F., and Barbacid, M. (1990). The trkB tyrosine protein kinase gene codes for a second neurogenic receptor that lacks the catalytic kinase domain. Cell 67, 647656. Klein, R., Jing, S., Nanduri, V., O’Rourke, E., and Barbacid, M. (1991). The trk proto-oncogene encodes a receptor for nerve growth factor. Cell 65, 169-I 97. Leibrock, J., Lottspeich, F., Hohn, A., Hofer, M., Hengerer, B., Masiakowski, P., Thoenen, H., and Barde, Y.-A. (1969). Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341,149152. Lin, L.-F. H., Mismer, D., Lile, J. D., Armes, L. G., Butler, E. T., Ill, Vannice, J. L., and Collins, F. (1969). Purification, cloning, and expression of ciliary neurotrophic factor. Science 246, 1023-1025. Lindsay, R. M., and Harmar, expression of neuropeptide 337,362-364.

A. J. (1969). Nerve growth factor regulates genes in adult sensory neurons. Nature

Maisonpierre, P. C., 6elluscio, L., Squinto, S., Ip, N. Y., Furth, M. E., Lindsay, R. M., and Yancopoulos, G. D. (199Oa). Neurotrophin-3: a neurotrophic factor related to NGF and BDNF. Science 247, 14461451. Maisonpierre, P. C., Belluscio, L., Friedman, B., Alderson, R. F., Wiegand, S. J., Furth, M. E., Lindsay, R. M.. and Yancopoulos, G. D. (1990b). NT-3, BDNF, and NGF in the developing rat nervous system: parallel as well as reciprocal patterns of expression. Neuron 5, 501509. Maisonpierre, P. C., Le Beau, M. M., Espinosa, R., Ill, Ip, N. Y., Belluscio, L.. De La Monte, S. M., Squinto, S. P., Furth, M. E., and Yancopoulos, G. D. (1991). Human and rat brain-derived neurotrophic factor and neurotrophin-3: gene structures, distributions, and chromosomal localizations. Genomics, in press. Martin-Zanca, D., Oskam, R., Mitra, G.. Copeland, T., and Barbacid, M. (1969). Molecular and biochemical characterization of the human frk proto-oncogene. Mol. Cell. Biol. 9, 24-33. Martin-Zanca, D., Barbacid, M., and Parada, L. F. (1990). Expression of the MC protooncogene is restricted to the sensory cranial and spinal ganglia of neural crest origin in mouse development. Genes Dev. 4, 663-694.

E. M. (1989).

Neurotrophic

molecules.

Squinto, S. P., Aldrich, T. A., Lindsay, R. M., Morrissey, D. M., Panayotatos, N., Bianco, S. M., Furth, M. E., and Yancopoulos, G. D. (1990). Identification of functional receptors for ciliary neurotrophic factor on neuronal cell lines and primary neurons. Neuron 5, 757-766. Stockli, K. A., Lottspeich, F., Sendtner, M., Masiakowski, P., Carroll, P., Gotz, R., Lindholm, D., and Thoenen, H. (1969). Molecular cloning, expression and regional distribution of rat ciliary neurotrophic factor. Nature 342, 920-923. Sutter, A., Riopelle, R. J., Harris-Warrick, R. M., and Shooter, E. M (1979a). Nerve growth factor receptors. Characterization of two distinct classes of binding sites on chick embryo sensory ganglia cells. J. BIoI. Chem. 254,5972-5962. Sutter, A., Riopelle, R. J., Harris-Warrick, (1979b). The heterogeneity of nerve growth membrane Signalling, M. Bitensky, R. J. C. F. Fox, eds. (New York: Alan R. Liss,

R. M., and Shooter, E. M. factor receptors. In TransCollier, D. F. Steiner, and Inc.), pp. 659-667.

Yancopoulos, G. D., Nisen, P. D., Tesfaye, A., Kohl, N. E., Goldfarb, M. P., and Alt, F. W. (1965). N-myc can cooperate with ras to transform normal cells in culture. Proc. Natl. Acad. Sci. USA 82, 5455-5459. Zimmerman, A., Sutter, A., Samuelson, J., and Shooter, E. M. (1978) A serological assay for the detection of cell surface receptors for nerve growth factor. J. Supramol. Struct. 9, 351-361.