The Journal of Neuroscience, July 1995, 15(7): 5420-5427

Synergy between Growth Factors and Transmitters Required for Catecholamine Differentiation in Brain Neurons

Xinyu Du and Lorraine lacovitti Institute of Neuroscience, Department of Neurology, Hahnemann University, Philadelphia, Pennsylvania 19102

The phenotypically plastic neurons of the embryonic rin) is necessary to trigger novel expression of the normally mouse striatum were used to explore mechanisms of cat- quiescent TH gene in non-CA neuronsof the striatum (Du et echolamine differentiation in culture. De nova transcription al., 1994). Whether aFGF and its partner substanceare also re- and translation of the CA biosynthetic enzyme, tyrosine hy- sponsiblefor signalingthe constitutive expressionof TH in CA droxylase (TH), was induced in striatal neurons exposed, neuronsis not yet clear. Consistentwith the latter notion, how- simultaneously or sequentially, to the growth factor, acidic ever, is the emergenceof TH in differentiating CA neurons fibroblast growth factor (aFGF) and a catecholamine. Al- (Specht et al., 1981) simultaneously with the appearanceof though dopamine was the most potent aFGF partner (EDSO aFGF in the developing brainstem (Ferrari et al., 1989; Engele = 4 PM), a number of substances, including dopamine (Dl) and Bohn, 1991; Fu et al., 1991; Schniirch and Risau, 1991; receptor agonists, f3-adrenoceptor agonists, and dopamine Wilcox and Unnerstall, 1991; Nurcombe et al., 1993). The pos- uptake inhibitors also trigger TH induction when accom- sibility is thus raised that CA neuronsharbor or have local ac- panied by aFGF. However, since none of the receptor an- cess to all of the agents necessaryfor their own biochemical tagonists nor transport blockers tested could inhibit do- differentiation, including, aFGF and its partner molecule. Sup- pamine’s action, the mechanism remains obscure. Struc- porting this propositionis our preliminary finding that an extract ture-activity analysis suggests that effective aFGF part- of CA neurons,derived from adult substantianigra (SN) tissue, ners all contain an amine group separated from a catechol is indeed capableof inducing TH expressionin cultured striatal nucleus by two carbons. Thus, TH expression can be nov- neurons.SN neurons,in addition to their supply of endogenous elly induced by the synergistic interaction of aFGF, and to aFGF (Engele and Bohn, 1991; Schniirch and Risau, 1991; Wil- a lesser extent basic FGF, and a variety of CA-containing cox and Unnerstall, 1991; Nurcombe et al., 1993), also contain partner molecules. We speculate that a similar association high concentrations of CA neurotransmitters,conceivably im- between growth factor and transmitter may be required in portant in aFGF activation. Since biogenic aminesare thought development for the differentiation of a CA phenotype in to subservea number of critical functions in fetal development brain neurons. (Coyle, 1972; Lauder, 1988; Mattson, 1988; Meier et al., 1991) [Key words: fibroblast growth factor, dopamine, tyrosine their potential participation in the CA differentiation processand hydroxylase, striatal neurons, tissue culture, catechol- probablemechanism of action were explored here. We will show amine differentiation] that induction of the CA-specific gene, TH, is mediated by a novel mechanismwhich requiresthe convergent actions of both Although the differentiation of the nervous system into neurons aFGF and, paradoxically, the enzyme’s catalytic end products, of specific biochemical classremains an elusory process,clues the CAs. That neurotransmittersmight work coordinately with as to the underlying mechanismsare now emerging.In the PNS, growth factors to initiate (Iacovitti et al., 1989; Iacovitti, 1991) studieshave indicated that individual growth factors act as crit- and subsequentlymodulate (DeVitry et al., 1991; Iacovitti et al., ical determinantsof transmitter type (Saadatand Thoenen, 1986; 1992; Louis et al., 1993; Magal et al., 1993; Iacovitti, 1994) Adler, 1989; Wong and Kessler, 1989; Yamamori et al., 1989; their own synthesis could provide an exciting new model for Rohrer, 1992; Howard and Gershon, 1993; Rao and Landis, autoregulationduring development and thereafter. The possible 1993). In the brain, however, the initiation of neurotransmitter- mechanismsthrough which various CAs might achieve these specific genesappears to involve more complex mechanisms, effects is discussed. requiring the obligatory interactionsof multiple signal molecules (Du et al., 1994). Although growth factors play a prominentrole, Materials and Methods their cooperation with auxiliary agents is essential.Thus, ex- Tissue culture. PregnantSwiss white mice were purchased from Taconic posure in culture both to the growth factor, aFGE and an addi- Lab Animalsat 9 d gestation2 12 hr fertilization day = embryonic day (E) 0. Pregnantdams were anesthetized with pentobarbitalon ges- tional unidentified muscle agent (demonstratednot to be hepa- tational day 13 and the embryos removed. The developing striatum was isolatedfrom the remainderof the brain as describedpreviously (Ia- Received Jan. 17, 1995; revised Mar. 16, 1995; accepted Mar. 21, 1995. covitti et al., 1989; Du et al., 1994). After removal of the meninges, We are grateful to Ms. Natalie Stull for her excellent technical assistance in tissue was minced(0.5 mm pieces)and incubatedin Ca2+/MgZ+-free all aspects of the project. This work was supported by NIH NS24204 and a Hanks’ balanced salt solution (CMF-HBSS) for 8 min at 37°C in a research gift from Amgen, Inc. clinical rotator (40 rpm). The incubation mixture was replaced with a Correspondence should be addressed to Lorraine Iacovitti, Institute of Neu- 0.01% trypsin solution (in CMF-HBSS), and incubated for an additional roscience, Department of Neurology, Hahnemann University, Broad and Vine 8 mitt, rinsed twice in Leibovitz’s medium (L-15), and placed in culture Streets, Philadelphia, PA 19102. medium containing Dulbecco’s minimum essential medium, 10% fetal Copyright 0 1995 Society for Neuroscience 0270-6474/95/155420-08$05.00/O calf serum(Irvine Scientific) glucose(6 mg/ml), glutamine(204 kg/ The Journal of Neuroscience, July 1995, 75(7) 5421 ml), and penicillin/streptomycin (100 U/ml). Cells were dissociated by Table 1. Effects of various neurotransmitter-related substances trituration through a reduced bore glass pipette and plated onto glass on TH induction in striatal neurons grown in the absence or Lab Tek culture dishes coated with polymerized polyornithine (pre- coated for 48 hr at room temperature at a concentration of 0.1 mg/ml presence of aFGF in culture. in 15 mM borate buffer, pH 8.4) and then rinsed with water and air dried 2-4 min. The cellular plating density was approximately 0.5-l X PercentageTH- 10” cells/dish. After a 1 d stabilization period in standard media, cultures positive neurons were incubated overnight in defined media (Bottenstein and Sato, 1979) Without With containing test reagents. The following day cultures were fixed and Treatment aFGF aFGF ED,,* processed for the immunocytochemical localization of TH unless oth- erwise stated. Control ND Norepinephrine ND 43 + 6 25 FM the immunoperoxidase/ABC method of staining (Elite Vectakit). Except where stated otherwise, TH induction was determined by counting pos- Epinephrine ND 37 + 5 85 pM itively stained cells in 50% off the microscopic fields on the culture Carbacol ND <5 dish. This was accomplished with the aid of an eyepiece reticule used Nicotine ND <5 at a 10X magnification. To eliminate the possibility that brainstem nor- Serotonin ND <5 adrenergic neurons were inadvertently included in dissected brain tissue, cultures of midbrain neurons were screened with antibodies to the nor- GABA ND <5 adrenergic enzyme dopamine-P-hydroxylase. At no time were stained Glutamate ND <5 midbrain neurons observed in these cultures. El3 mouse striatal neurons were established in culture as described previously Pharmacological studies. El 3 mouse striatal neurons were estab- (Iacovitti et al., 1989). Cultures were stabilized one day prior to the addition lished in culture as described previously (Iacovitti et al., 1989). Cultures of various neurotransmitter substances to the media (at a concentration range were stabilized 1 d prior to the addition of various reagents to the media of I-800 p,M). Some of the cultures were simultaneously treated with exoge- (see Table 2 for listing). To examine the role of dopamine receptors nous aFGF (10 “g/ml). The following day, all cultures were fixed in 4% para- (Dl-D5) and adrenoceptors (a and p) in TH induction, cultures were formaldehyde, immunocytochemically stained for the localization of TH, and simultaneously treated with test compounds (at a concentration range scored for TH induction as described previously (Iacovitti et al., 1989). The of I-200 by) and aFGF (10 rig/ml). In other cultures, we evaluated the results are expressed as the percentage of total neurons which stained positively with antibodies to TH. Each value represents the mean + the SD for triplicate ability of .test compounds (i.e.: receptor antagonists or transport inhib- cultures in 2 separate platings. itors: I-200 U.M) to inhibit TH induced bv aFGF (10 rig/ml) + CA (eitherDA at’lO’+M or NE at 50-100p,t+*Culture media-wai supple- ND, not detectable. mented with test reagents 2 hr prior to the addition of TH-inducing * ED,,, values were calculated for cultures co-treated with aFGF (IO “g/ml) agents. The day after the various treatments, all cultures were fixed in and l-200 pM neurotransmitter. 4% paraformaldehyde,immunocytochemically stained for the localiza- tion of TH, and scored for TH induction as described previously (Ia- covitti et al., 1989). The results are expressed as the percentage of total vealed that other CA-containing compounds,such as, L-dopa, neurons which stained positively with antibodies to TH. norepinephrine and epinephrine were also effective, albeit at Statistical analysis. Data were statistically analyzed by one-way anal- higher concentrations(ED,, = 25-85 PM). The role of catecho- ysis of variance. When P < 0.05, then theF test was followed by the two-tailed Student’s t test to compare the statistical significance between lamines in this effect may be important for explaining the low control and experimental groups. Differences were considered signifi- level TH induction sometimesobserved when aFGF was added cant only when the P value was less than 0.05. to the media in the absenceof exogenouspartner molecules(Du Chemical reagents. All transmitter-related compounds, cyclohexam- et al., 1994). In all casesof basalTH induction, scatteredTH- ide, and cY-amanitinwere purchasedfrom SigmaChemical (St. Louis, immunoreactive neurons were observed, even in control (un- MO). All receptor agonist and antagonists were purchased from RBI (Natick, MA). Growth factors were acquired from the following treated) cultures. It is conceivable that the presencein some sources: aFGF, bFGF, and FGF-7 were kind gifts of Amgen; FGF-6 was platings of thesespurious TH neuronsresulted in the production kindly suppliedby Dr. E Coulier(Marseille, France); fiGF was a gift of low levels of endogenousL-dopa or dopamine which then from Dr. E. M. Johnson (St. Louis. MO): CNTF was kindlv suoolied partially induced TH expressionin other neuronswhen aFGF by Scios Nova Inc.; EGE‘LIE IL1 ‘and TGFP were purchaskd f;dm R and D Systems. All reagents used in culture were obtained from GIB- was added exogenously. co. In contrast to CAs, when cultures of striatal neurons were incubated with various concentrationsof monoamineprecursors Results or their metabolites (1 .O PM to 200 PM of tyrosine, tyramine, Induction of TH by aFGF and CAs homovanillic acid and 3,Cdihydroxyphenylacetic acid) alone or The phenotypically plastic neurons of the El3 mouse striatum in combination with aFGF (10 rig/ml), no TH induction was were chosen as the model system for these studies since their observed (data not shown). neurotransmitter profile can be permanently altered by exposure to differentiative cues in culture (Iacovitti et al., 1989; Iacovitti, Studies on the mechanism of CA action 1991; Du et al., 1994). As part of our initial screening,we tested We next investigatedwhether CAs mediatedtheir cellular effects the capacity for a variety of neurotransmittersand related com- through either dopaminergicladrenergicmembrane receptors or pounds to induce TH in striatal neuronswhen incubated alone by direct uptake into the cell via transporter mediatedsystems. or in combinationwith aFGE As seenin Table 1, administration Summarizedin Table 2 is a list of receptor agonist and antag- of neurotransmittersubstances or aFGF alone had no apparent onist compoundsand transport blockers which were testedin an effect on TH expression.In contrast, the combinedtreatment of attempt to either stimulate or inhibit CA assistedTH induction aFGF with transmitters,specifically CA transmitters,resulted in by aFGE All reagentswere first evaluated for TH-inducing ca- a striking induction in TH in striatal neurons. Dopamine was pacity when administeredalone. In no casewas TH inducedby particularly effective as an aFGF partner (ED,, = 4 FM), initi- treatment solely with a receptor agonist, antagonist or uptake ating the expressionof TH in the majority of cultured striatal blocker (data not shown). To begin to understandthe role of neurons(Fig. 1). However, dose-responsestudies (Fig. 2) re- dopaminereceptors (Dl-D5) and/or adrenoceptors(a and p) in 5422 Du and lacovitti * Dopamine and aFGF Together Induce Tyrosine Hydroxylase

Figure 1. Immunocytochemical localization of TH in cultured striatal neurons. Neurons were established in culture (Iacovitti et al., 1989; Iacovitti, 1991) 1 d prior to incubation in control media (A) or in media supplemented with 10 nglml aFGF (B), 10 pM dopamine (C); or 10 rig/ml aFGF + 10 p,M dopamine (D). The following day, cultures were fixed and TH was immunocytochemically localized. Note that dopamine, in collaboration with aFGF, produced a striking induction of TH in many of cultured striatal neurons. mediating TH induction, cultures were simultaneouslytreated fold excess concentration; 400 X the IC,,) was able to block with a maximal inducing doseof aFGF (10 rig/ml) and selective induction by aFGF and dopamine(Table 2). Moreover, selective CA receptor agonists/antagonists(see Table 2 listing) at a con- inhibitors of GABA uptake did not prevent TH induction, sug- centrationrange of l-200 pM. We found that the DUD5 receptor gesting that the GABA transporteris not responsiblefor taking agonistSKF 38393, in all rotameric conformations(+, -, + ), up CAs in theseexperiments. Likewise dopaminetransport in- could successfully substitute for dopamine, inducing TH ex- hibitors could not block, though several (mazindol, GYKI pressionin 60% of those striatal neuronssimultaneously treated 52895) paradoxically mimicked the effects of dopamine, pro- with aFGE However, pre- and cotreatment with known Dl re- ducing 40% TH induction when co-incubated with aFGE More- ceptor antagonists(I-200 pM) were unable to block this effect. over, in the case of mazindol, the level of induction could be Neither SCH 23390, apomorphine,haloperidol could inhibit TH further increasedby the addition of dopamineas well as aFGE induced by aFGF (10 rig/ml) + dopamine (10 FM) or SKF Although it is possible that mazindol, by blocking reuptake of 38393 (10 pM; data not shown). Other dopaminereceptor (D2, dopamine,increases extracellular levels of transmitter sufficient- D3, D4) agonistsand antagonistsneither stimulatednor inhibited ly to partner aFGE this does not explain its ability to enhance TH expressionin striatal neurons.We next tested the potential TH induction when co-incubated with a maximally inducing role of classic adrenergicreceptors in TH induction. The B-ad- dose of dopamine.The additive effects of mazindol and dopa- renergic receptor agonist, isoproterenol,similar to NE (reported mine imply that both agentsmay partner aFGF but their mech- above), was a potent aFGF partner, inducing TH immunoreac- anismsof action may differ. tivity in the majority of striatal neurons(Table 2). Again, how- ever, the B-adrenergicantagonist propranolol was incapable of CA activating effect is maximal for aFGF compared to other preventing induction by aFGF and NE or Isoproterenol. growth factors Since it is possiblethat CAs do not work at the membrane, In addition to aFGE ongoing studies(Du et al., 1995) suggest but rather exert their effects intracellularly following uptake via that other dopaminedifferentiation factors, suchas brain-derived a neurotransportsystem, we attemptedwith known DA uptake growth factor (BDNF) (but not glial-derived growth factor), may blockers (mazindol, nomifensine, etc.) or GABA uptake block- also possesssome ability to induce TH expressionin the pres- ers (nipecotic acid; 1-amino- 1 -cyclohexane carboxylic acid, ence of partner molecules(30% induction). We therefore tested ACCA) to inhibit TH induction. As we had seenwith the re- whether other neuralgrowth factors could initiate TH when part- ceptor antagonists,no uptake compound tested (even at a 20- neredby CA. Striatal cultureswere incubatedovernight in media The Journal of Neuroscience, July 1995, f5(7) 5423

100 --)- DA + aFGF was added first and then replacedwith (k)SKF 38393 or vice versa (data not shown). However, aFGF required a longer ex- ...... 0 ...... f.,E +aFGF posuretime (4 hr) than did (?)SKF 38393 (30 min) for initi- ating the effect. When both reagentswere addedsimultaneously ____ +.. L-Dopa + aFGF 75 to the culture, increasedTH expressionwas also observedwithin m i --A---- Epinevhrine + aFGF 4 hr of cotreatment. The fact that both aFGF and the dopamine receptor agonistneed not be incubatedtogether suggests that the effect occurs independentof conformational changesresulting from their direct molecular interaction. Instead, it appearsthat convergenceof aFGF and its partner occursinside the cell, pos- sibly along their intracellular signalingpathways. De novo transcription and translation are required for TH induction by aFGF and dopamine As suggestedby our earlier studies (Iacovitti et al., 1989; Ia- covitti, 1991 ), the induced appearanceof immunoreactiveTH in striatal neurons,required both de novo transcription and trans- lation. Thus, inhibition of protein synthesiswith cyclohexamide decreasedin a dose-dependentmanner the number of TH im- d v- munoreactiveneurons elicited by dopamineand aFGF treatment; inhibiting over 90% of TH induction at the maximum dosetest- Concentration (uM) ed (1000 rig/ml) (Fig. 4). Preincubation of neurons with the Figure 2. Dose dependency of the effect of catecholamines on aFGF- RNA polymeraseinhibitor a-amanitin, at a doserange (4 kg/ induced TH expression in El3 cultured striatal neurons. One day after ml) which totally inhibits RNA polymeraseII (Roeder, 1976), plating, cultures were fed media supplemented with 10 rig/ml aFGF and reduced by more than 50% the number of TH immunoreactive various concentrations (0.3-800 PM) of dopamine (square),norepi- neurons(Fig. 4). Higher concentrations(400 pg/ml) of amanitin nephrine (NE) (circle), L-dopa (diamond), or epinephrine (triangle). The following day, cultures were fixed and evaluated for TH induction. which inhibit RNA polymeraseIII as well (Stott, 1991), pre- Values are expressed as the percentage of TH positive neurons in the vented 90% of the usual induction of TH. culture. Each value represents the mean -t SD of four determinations for two separate platings. Discussion The results of these studiessuggest that, in mammalianbrain, as in simpler invertebrate systems(Karin, 1989; Foulkes et al., containing aFGF, bFGF, FGF-6, FGF-7, epidermalgrowth factor 1991; Lamb and McKnight, 1991; Bowers, 1994), geneexpres- (EGF), NGE ciliary neurotrophic factor (CNTF), leukemia in- sion is controlled by the synergisticand competitive interactions hibitory factor (LIF), transforming growth factor B (TGFB), or of severalregulatory factors. Thus, the growth factor aFGF, and interleukin 1 (ILI) alone or in combination with 10 pM dopa- to a lesserextent bFGE working in concert with a partner mol- mine. The following day TH induction was evaluated immu- ecule, shown here to be a CA neurotransmitter,will trigger the nocytochemically. Of the growth factors tested, only aFGF, and novel expressionof TH in non-CA neurons. This expression to a lesserextent, its isoform basic FGF (bFGF) could induce requires both de novo transcription of the TH gene and trans- TH expressionwhen dopaminewas included in the incubation lation of the mRNA into newly synthesizedprotein. media (Fig. 3). Although the maximum level of induction was lower for bFGF than aFGE both growth factors generatednearly Mechanism of TH induction identical time courseswhen incubated with dopamine:TH ex- The mechanismsthrough which the cultured striatal neuron pressionwas first evident at 6 hr, peakedby 12 hr and remained achievesthese actions remain elusory. It is presumedbut not yet elevated even 108 hr following treatment (Fig. 3). proven that aFGF exerts its effects by binding to one or more of the four high affinity tyrosine kinase-linkedmembrane recep- Synergy doesnot require co-incubation of aFGF and CAs tors (Donne et al., 1991). CAs, on the other hand, may mediate In an attempt to localize the synergy betweenaFGF and its part- their cellular actionsthrough severalroutes. First, CA membrane ner molecule, TH induction was compared after the simulta- receptors, including dopamine (DUD2) receptors (Maus et al., neousor sequentialincubation of inducing agents.TH induction 1989; Premont et al., 1983) as well as (Y- (Weiss et al., 1987) was compared in cultures incubated simultaneouslywith aFGF and B- (Van Vliet et al., 1991) adrenoceptorspresent on cultured (10 rig/ml) and (+)SKF 38393 (100 pM) and those which re- striatal neurons might activate second messengersystems to ceived successivetreatments with the individual reagents.In the bring about TH induction. Indeed, in our studies,agonists of the former group, both reagentswere added together for 0.254.0 DUD5 dopaminereceptor and the B-adrenergicreceptor mim- hr, media was then removed, cultures were rinsed twice for 5 icked the TH inductive effects of dopamine and NE, respective- min each and reincubatedfor the remainingtime in defined me- ly. However, the fact that specific antagonistsof thesesites did dia (which had not been supplementedwith added agents). In not inhibit their action (even in concentrations400X the IC,,) the latter group, cultureswere treatedfor various periodsof time suggeststhat thesereceptors are not the mediatorsof TH induc- (0.254.0 hr) with aFGF or (?)SKF 38393, rinsed as above and tion. then re-incubated with the reciprocal agent. The next day, cul- Although there is no evidence for the existence of a dopamine tureswere fixed and stainedwith TH antibodies.After sequential transporter (Shimadaet al., 1992; Cerruti et al., 1993; Grant et treatment, induction was observed regardlessof whether aFGF al., 1994)on GABAergic neuronsof the adult striatum,we none- 5424 Du and lacovitti * Dopamine and aFGF Together Induce Tyrosine Hydroxylase

Table 2. Effects of various receptor agonists, antagonists, or uptake inhibitors on TH induction.

TH induction Treatment Site of action Main action +aFGF* +aFGF + DA** Dopamine receptors R (+) SKF 38393 Dl/D5 agonist +++ ND S (-) SKF 38393 Dl/D5 agonist +++ ND RS (?) SKF 38393 Dl/D5 agonist +++ ND R (+) 6-bromoAPB Dl/D5 agonist + ND R (+) SCH 23390 Dl/D5 antagonist ~ +++ Quinpirole D2/D3/D4 agonist - ND Spiperone D2/D3/D4 antagonist - +++ Apomorphine Dl/D2 agonist + +++ Halperidol Dl/D2 antagonist ND +++ SCH-23390 + Spiperone D l/D2/D3/D4 antagonist ND +++ +aFGF +aFGF + NE’ Adrenoceptors Phenylephrine alpha 1 agonist - ND Prazosin alpha 1 antagonist ND ++ Clonidine alpha 2 agonist ND Yohimbine alpha 2 antagonist - ++ Yohimbine + Prazosin alpha l/2 antagonist ND ++ Isoproterenol beta agonist ++ ND Propranolol beta antagonist ~ ++ + aFGF +aFGF + DA Uptake sites Mazindol DA transporter blockade + + ++++ Nomifensine DA transporter blockade ~ +++ GYKI 52895 DA transporter blockade + + +++ GBR-I 2935 DA transporter blockade + +++ WIN 35,428 DA transporter blockade - +++ Indatraline DA,NE,S-HT blockade ~ +++ ACCA GABA transporter blockade - +++ RS (2 ) Nipecotic acid GABA transporter blockade - +++

El3 mouse striatal neurons were established in culture as described previously (Iacovitti et al., 1989). Cultures stabilized one day prior to the addition of various reagents to the media. The ability of compounds to substitute for CA in TH induction was evaluated by co-treatment of cultures with l-200 pM of the receptor agonist, antagonist, or uptake inhibitor and 10 rig/ml of aFGF (left column). The capacity to block TH induction by aFGF + CA (either dopamine [DA] at 10 p,M or NE at 50P100 PM) was assessed by pretreatment of the cultures for 2 hr with the test compound before the further addition of aFGF (10 rig/ml) + the appropriate CA (10 PM) to the media (right column). On the day following the various treatments, all cultures were fixed in 4% paraformaldehyde, immuno- cytochemically stained for the localization of TH, and scored for TH induction as described previously (Iacovitti et al., 1989). The results are expressed as the percentage of total neurons which stained positively with antibodies to TH. Each + represents approximately 20% TH immunoreactive neurons (i.e., +++ = 60%); -, no induction in TH expression over control levels; ND, not determined. Controls: * aFGF (10 rig/ml) only yields no TH induction; ** aFGF (10 rig/ml) + DA (10 PM) yields ++ + TH induction; ! aFGF (10 rig/ml) + NE (50 FM) yields ++ TH induction. theless tested whether dopamine was gaining access to the cell but expression of the former is lost by adulthood. Precedent for via an uptake mechanism. When agents known for their ability the coexistence of GABA and DA transporters on the same cell to block dopamine transporter activity were tested, they did not has indeed been reported previously (Bonanno and Raiteri, prevent the induction of TH by aFGF and dopamine. In fact, 1987). Alternatively, it is possible that the embryonic GABA several dopamine uptake inhibitors were instead inducers of TH transporter possesses recognition and/or translocation sites when co-incubated with growth factor. This group of apparently which are less discriminating than the adult, permitting uptake disparate aFGF partners (Dl/D5 agonists, P-adrenergic agonists, of dopamine and other related catechols. In view of the fact that dopamine uptake inhibitors) share several distinguishing bio- it is now believed that the gene family of neurotransporters is chemical features; all members contain an amine group separat- far larger than originally suspected (i.e., “orphan” transporters ed from a catechol nucleus by two carbons. The fact that these with no known function have been identified) (Uhl, 1987), an chemical traits are virtually identical to those required for opti- entirely new neurotransporter may account for our results. Re- mal uptake by the dopamine transporter (Meiergerd and Schenk, gardless of whether striatal neurons possess an embryonic form 1994) raises several possibilities. First, it is conceivable that em- of the DA or GABA transporter or a new transporter, our results bryonic striatal neurons contain the dopamine neurotransporter make clear that this site can not be blocked by the usual inhib- in addition to the usual GABA transporter (Radian et al., 1990) itors of dopamine or GABA uptake. The Journal of Neuroscience, July 1995, 15(7) 5425

mm.- -.w T T T o- DA --4--- aFGF

-.-. m.-. bFGF I’\ ‘\ ‘\ ‘\ e DA + aFGF DA + bFGF

0 48 72 96 120 Treatment Time (hr.)

Figure 3. Time course of TH expression in striatal neurons incubated in media containing 1 mg/ml of BSA (triar.gle), 10 p,M dopamine (DA, circle), 10rig/ml aFGF (opendiamond), 10rig/ml bFGF (square), DA + aFGF(shaded diamond), DA + bFGF (shadedsquare). Cultureswere fixed at varioustime intervalsup to 108hr andTH inductionwas evaluated immunocytochemically. Values are expressed as the percentageof TH positiveneurons in the culture.Each value representsthe mean2 SD of threeto six determinationsfor two separateplatings.

Finally, we can not overlook the possibility that, since TH 1 d before TH first appearsin the CNS (Specht et al., 1981). induction is not inhibited by classicreceptor antagonistsor trans- An alternative sourceof CAs are those derived externally from port blockers and can be initiated by both D and L forms of the the maternal circulation which crossthe placenta to freely per- receptor agonist SKF 38393, dopamineutilizes a novel mecha- meatethe embryonic brain (Saundersand Mollgard, 1984; Los- nism to mediate its effects on striatal neurons. Since the by- sinsky et al., 1986). This putative role for CAs in differentiation products of lipid peroxidation are known to increasegene tran- is indeed consistentwith the widely held view that neurotrans- scription and stimulate biochemical differentiation in a variety mitters mediate many critical functions in fetal development of developing tissues(Allen, 1991), one intriguing possibility is (Coyle, 1972; Lauder, 1988; Mattson, 1988; Meier et al., 1991), that catecholsinduce TH by increasinglipid peroxidation of the including the stimulation of differentiation factor production cell membrane.Resolution of theseissues awaits further inves- (Habeckerand Landis, 1994). Nonetheless,we can not eliminate tigation. the possibility that CAs are not the physiologically relevant part- Relevance to CA differentiation nersduring developmentbut are merely mimicking their effects by activating signalingpathways in commonwith thoserequired Whether these in vitro results on the non-CA neurons of the for TH induction. striatum are indicative of an in vivo role for aFGF and catecho- The induction of TH elicited by aFGF and dopamine was lamines in initiating TH gene expressionin CA neuronsis not mechanistically similar to that seenpreviously with L6 muscle yet known. Certainly, aFGF is found locally in the brainstem (Ferrari et al., 1989; Engele and Bohn, 1991; Fu et al., 1991; extract (Iacovitti et al., 1989; Iacovitti, 1991), requiring both de Schniirch and Risau; 1991; Wilcox and Unnerstall, 1991; Nur- IZOVOtranscription and translation.However, musclecells, which combe et al., 1993) during the period when CA neuronsfirst manufacture their own aFGF, do not synthesizeor take up do- differentiate (Specht et al., 1981). However, the presenceof oth- pamine or related neurotransmittercompounds. It is, therefore, er dopaminedifferentiation factors, such as BDNE may also unlikely that the partner molecule present in L6 cells is one of prove to be physiologically important for TH activation during the CA-containing compoundsidentified here. Implicit in this development (Du et al., 1995). Regardlessof the growth factor, conclusionis the existence of multiple activators of aFGE some activation by partner molecules, such as CAs, appearsto be a found iri neurons(i.e., CAs), and at least one other present in necessarystep in TH gene initiation. The requirement for en- muscle. Although the required participation of a partner mole- dogenousCAs, however, posesan obvious paradox for neurons cule may be useful for restricting the effects of universal growth attempting to first initiate CA synthesis.Consequently, biogenic substancessuch as aFGF, the availability of several partners aminesmust originate extraneuronally if they are to act as aFGF could provide the versatility neededto inducethe TH geneunder partners during differentiation. One potential internal source of a variety of different conditions. Conceivably, these different CAs might be cells of the PNS (i.e., sympathetic neurons,ad- partner moleculesmight achieve their effects via activation of renal medullary cells) which acquire the capacity to synthesize one or more of the many important transcription binding sites CAs by El 1.5 in the rat (Teitelmanet al., 1981), approximately already identified on the TH gene(Lewis and Chikaraishi, 1987; 5426 Du and lacovitti - Dopamine and aFGF Together Induce Tyrosine Hydroxylase

expression of a quiescent TH gene are also involved in its rou- tine regulation. Although CA transmitters have long been con- sidered feedback inhibitors of the TH enzyme (Nagatsu et al., 1964; Spector et al., 1967; Weiner et al., 1972; Zigmond et al., 1989), their possible role when combined with the growth factor, aFGF, as feedback inducers of TH has not been explored pre- viously. Thus, as in other systems (Mani et al., 1994), the cross talk between growth factors and neurotransmitters may be es- sential for proper regulation of TH at the transcriptional and translational levels. References Adler JE (1989) Neuronal aggregation and neurotransmitter regulation: partial purification and characterization of a membrane-derived factor. Int J Dev Neurosci 7:533-538. Allen RG (1991) Oxygen-reactive species and antioxidant responses during development: the metabolic paradox of cellular differentiation. Proc Sot Exp Biol Med 196: 117-129. Black IB (1993) Environmental regulation of brain trophic interactions. Int J Dev Neurosci 11:403-410. Bonanno G, Raiteri M (1987a) Coexistence of carriers for dopamine and GABA uptake on a same nerve terminal in the rat brain. Br J Pharmacol 91:237-243. CHX (rig/ml) Bottenstein JE, Sato GH (1979) Growth of rat neuroblastoma cell line in serum-free supplemented medium. Proc Nat1 Acad Sci USA 76: 514-517. Bowers CW (1994) Superfluous neurotransmitters? Trends Neurosci 17:315-320. Carroll JM, Evinger MJ, Goodman HM, Joh TH (1991) Differential and coordinate regulation of TH and PNMT MRNAs in chromaffin cell cultures by second messenger system activation and steroid treat- TT T ment. J Mol Neurosci 3:75-83. Cerruti C, Walther DM, Kuhart MJ, Uhl GR (1993) Dopamine trans- porter mRNA expression is intense in rat midbrain neurons and mod- est outside midbrain. Mol Brain Res 18:181-186. Coyle JT (1972) Tyrosine hydroxylase in rat-brain-cofactor require- ments, regional and subcellular distribution. Biochem Pharmacol 21: 1935-1944. DeVitry F, Hillion J, Catelon J, Thibault J, Benoliel JJ, Hamon M (1991) Dopamine increases the expression of tyrosine hydroxylase and aromatic amino acid decarboxylase in primary cultures of fetal neurons. Dev Brain Res 59:123-131. Donne CA, Jaye M, Schlessinger J (1991) Structural diversity and binding of FGF receptors. In-Annals of the New York Academy of Sciences. Vol 638. The fibroblast growth factor familv (Baird A. Klasbrun M, eds), ‘pp 161-166. New York: New York Academy of Sciences. Du XY, Stull ND, Iacovitti L (1994) Novel expression of the tyrosine hydroxylase gene requires both acidic fibroblast growth factor and an activator. J Neurosci 14:7688-7694. Du X, Stull ND, Iacovitti L (1995) Brain-derived neurotrophic factor works coordinately with partner molecules to initiate tyrosine hy- alpha-Amanitin @g/ml) droxylase expression in striatal neurons. Brain Res, in press. Engele J, Bohn MC (1991) The neurotrophic effects of fibroblast Figure 4. Effect of protein and RNA synthesis inhibitors on the TH growth factors on dopaminergic neurons in vitro are mediated by expression elicited by aFGF and dopamine. Neurons stabilized for 1 d mesencephalic glia. J Neurosci 11:3070-3078. in culture received 0.01-1000 rig/ml cyclohexamide (CHX) to inhibit Ferrari G, Minozzi MC, Toffano G, Leon A, Skaper SD (1989) Basic protein synthesis (A) or 0.25400 yg/ml a-amanitin to inhibit RNA fibroblast growth factor promotes the survival and development of polymerases (B) 2 hr before the addition of 10 FM dopamine, 10 ng/ mesencephalic neurons in culture. Dev Biol 133:140-147. ml aFGF to the culture media. The following day cultures were fixed Ferris RM, Tang FLM, Russell AV (1975) Effects of apomorphine in and stained for TH. Values are expressed as the percentage of TH pos- the uptake and release of catecholamines in crude synaptosomal prep- itive neurons in the culture. Each value represents the mean + SD of arations of rat striatum and hypothalamus. Biochem Pharmacol 24: four determinations. 1523-1527. Foulkes NS, Borrelli S, Sassone-Corsi P (1991) CREM gene: use of alternative DNA-binding domains generates multiple antagonists of Jones et al., 1988; Kemmler et al., 1989; Carroll et al., 1991; CAMP-induced transcription. Cell 64:739-749. - - Fung et al., 1992; Kim et al., 1993; Murre and Baltimore, 1993). Fu Y-A. Snirito R Yu Z-X. Biro S. Sasse J. Lei J. Ferrans VJ. Enstein Consistent with this view is the recent finding that tissue specific SE, da&ells W (1991) Acidic fibroblast growth factor in the de- expression of TH indeed requires an interaction between two veloping rat embryo. Cell Biol 114:1261-1273. Fung BP Yoon SO, Chikaraishi DM (1992) Sequences that direct the consensus binding sites on the gene (Yoon and Chikaraishi, rat tyrosine hydroxylase gene expression. J Neurochem 58:2044- 1992). 2052. It is not yet clear whether the same substances that first initiate Grant S, Revay R, Vaughan R, Freed C, Uhl G, Kuhar MJ (1994) The Journal of Neuroscience, July 1995, 15(7) 5427

Immunohistochemical localization of dopamine transporters in rat Nurcombe V, Ford MD, Wildschut JA, Barlett PF (1993) Development brain. Sot Neurosci Abstr 20:920. regulation of neural response to FGF-1 and FGF-2 by hepaian sulfate Habecker BA, Landis S C (1994) Noradrenergic regulation of cholin- proteoglycan. Science 260: 103-106. ergic differentiation. Science 264: 1602-1604. Premont J, Dauguet-de Montety M-C, Herbet A, Glowinski J, Bockaert Howard MJ, Gershon MD (1993) Role of growth factors in catechol- J, Prochiantz A (1983) Biogenic amines and adenosine-sensitive ad- aminergic expression by neural crest cells: in vitro effects of trans- enylate cyclases in primary cultures of striatal neurons. Brain Res forming growth factor beta,. Dev Dyn 196:1-10. 285:53-61. Iacovitti L (1991) Effects of a novel differentiation factor on the de- Radian R, Ottersen OP, Storm-Mathisen J, Caste1 M, Kanner BI (1990) velopment of catecholamine traits in noncatecholamine neurons from Immunocytochemical localization of the GABA transporter in rat various regions of the rat brain: studies in tissue culture. J Neurosci brain. J Neurosci 10: 1319-1330. 11:2403-2409. Rao MS, Landis SC (1993) Cell interactions that determine sympa- Iacovitti L (1994) Centrally active differentiation factors in the nervous thetic neuron transmitter phenotype and the neurokines that mediate system. Pharmacol Ther 62:373-383. them. J Neurobiol 24:215-232. Iacovitti L, Evinger MJ, Joh TH, Reis DJ (1989) A muscle-derived Roeder RG (1976) Eukaryotic nuclear RNA polymerases. In: RNA factor induces expression of a catecholamine phenotype in cultured polymerase (Losick R, Chamberlin M, eds), pp 285-329. Cold Spring rat cerebral cortex. J Neurosci 9:3529-3537. Harbor, NY: Cold Spring Harbor Laboratory. Iacovitti L, Evinger MJ, Stull ND (1992) Muscle-derived differentia- Rohrer H (1992) Cholinergic neuronal differentiation factors: evidence tion factor increases expression of the tyrosine hydroxylase gene and for the presence of both CNTF-like and non-CNTF-like factors in enzyme activity in cultured dopamine neurons from the rat midbrain. developing rat footpad. Development 114:689-698. Mol Brain Res 16:215-222. Saadat S, Thoenen H (1986) Selective induction of tyrosine hydroxy- Jones NC, Rigby PWJ, Ziff EB (1988) Trans.acting protein factor and lase by cell-cell contact bovine adrenal chromaffin cells is mimicked the regulation of eukaryotic transcription: lessons from studies on bv nlasma membranes. J Cell Biol 103:1991-1997. DNA tumor viruses. Genes Dev 2:267-281. Saunders NR, Mollgard K (1984) Development of the blood-brain bar- Karin M (1989) Complexities of gene regulation by CAMP. Trends rier. J Dev Physiol 6:45-57. Genet 5:65-67. Schniirch H, Risau W (1991) Differentiating and mature neurons ex- Kemmler I, Schreiber E, Muller MM, Matthias P, Schaffner W (1989) press the acidic fibroblast growth factor gene during chick neural Octamer transcription factor bind to two different sequence motifs of development. Development 111: 1143-l 154. the immunoglobulin heavy chain promotor. EMBO J 8:2001-2008. Shimada S, Kitayama S, Lin C-L, Pate1 A, Nanthakumar E, Gregor P, Kim KT, Park DH, Joh TH (1993) Parallel up-regulation of catechol- Kuhar M, Uhl G (1991) Cloning and expression of a cocaine-sen- amine biosynthetic enzymes by dexamethasone in PC12 cells. J Neu- sitive dopamine transporter complementary cDNA. Science 254:576- rochem 60:946-95 1. 577. Lamb P, McKnight SL (1991) Diversity and specificity in transcrip- Shimada S, Kitayama S, Walther D, Uhl G (1992) Dopamine trans- tional regulation: the benefits of heterotypic dimerization. Trends Bio- porter mRNA: dense expression in ventral midbrain neurons. Mol them 16:417427. Brain Res 13:359-362. Lauder J (1988) Neurotransmitters as morphogens. Prog Brain Res 73: Specht LA, Pickel VM, Joh TH, Reis DJ (1981) Light microscopic 365-387. - immunocytochemical localization of tyrosine hydroxilase in prenatal Lewis EJ, Chickaraishi DM (1987) Regulated expression of the tyro- rat brain. I. Earlv ontoaenv. J Comn Neurol 199:233-253. sine hydroxylase gene by epidermal growth factor. Mol Cell Biol Spector S, Gordonopioid receptors inn the rat striatum. Eur J Pharmacol Convergent pathways for steroid hormone- and neurotransmitter-in- 195:295-300. duced rat sexual behavior. Science 265: 1246-l 249. Weiner N, Cloutier G, Biur R, Pfeffer RI (1972) Modification of nor- Mattson M (1988) Neurotransmitters in the regulation of neuronal cy- epinephrine synthesis in intact tissue by drugs and during short-term toarchitecture. Brain Res Rev 13: 179-212. adrenergic nerve stimulation. Pharmacol Rev 24:203-221. Maus MP Bertrand S, Drouva R, Rasolonjanahary C, Kordon J, Glow- Weiss S, Kemp DE, Lenox RH, Ellis J (1987) a,-Adrenergic receptors inski J, Premont J, Enjalbert A (1989) Differential modulation of D, mediate inhibition of cyclic AMP production in neurons in primary and D, dopamine-sensitive adenylate cyclases by 17B-estradiol in culture. Brain Res 414:390-394. cultured striatal neurons and anterior pituitary cells. J Neurochem 52: Wilcox B, Unnerstall J (1991) Expression of acidic fibroblast growth 410-418. factor mRNA in the developing and adult rat brain. Neuron 6:397- McElvain JS, Schenk JO (1992) A multisubstrate mechanism of striatal 409. dopamine uptake and its inhibition by cocaine. Biochem Pharmacol Wong V, Kessler JA (1987) Solubilization of a membrane factor that 434:2189-2199. stimulates levels of substance P and choline acetyltransferase in sym- Meier E, Hertz L, Schousboe A (1991) Neurotransmitters as develop- pathetic neurons. Proc Nat1 Acad Sci USA 84:8726-8729. mental signals. Neurochem Int 19:1-15. Yamamori T, Fukada K, Aebersold R, Korsching S, Fann MJ, Patterson Meiergerd SM, Schenk JO (1994) Striatal transporter for dopamine: PH (1989) The cholinergic neuronal differentiation factor from heart catechol structure-activity studies and susceptibility to chemical cells is identical to leukemia inhibitory factor. Science 246: 1412- modification. J Neurochem 62:998-1008. 1416. Murre C, McCaw PS, Baltimore D (1989) A new DNA binding and Yoon SO, Chikaraishi DM (1992) Tissue-specific transcription of the dimerization motif in immunoglobulin enhancer binding, daughter- rat tyrosine hydroxylase gene requires synergy between and AP-1 less MyoD, and myc proteins. Cell 56:777-783. motif and an overlapping E box-containing dyad. Neuron 9:55-67. Nagatsu T, Levitt M, Udenfriend S (1964) Tyrosine hydroxylase: the Zigmond RE, Schwarzschild MA, Rittenhouse AR (1989) Acute reg- initial step in norepinephrine biosynthesis. J Biol Chem 239:2910- ulation of tyrosine hydroxylase by nerve activity and by neurotrans- 2917. mitters via phosphorylation. Annu Rev Neurosci 12:415-461.