[1] 23w Regulation of Vasoactive Intestinal Peptide Expression in Sympathetic Neurons in Culture and after Axotomy: The Role of Cholinergic Differentiation Factor /Leukemia Inhibitory Factor

[1] 141w Vasoactive intestinal peptide (VIP) expression increases in sympathetic neurons when they are grown in dissociated cell or explant cultures and when they are axotomized in vivo. In dissociated cell culture, the magnitude of the VIP increase was reduced when nonneuronal cells were removed and medium conditioned by ganglionic nonneuronal cells increased V I P in neuron-enriched cultures. Antiserum against cholinergic differentiation factor (also leukemia inhibitory factor; CDF /LIF), but not against ciliary neurotrophic factor, immunoprecipitated this activity. Medium conditioned by sympathetic gan-glion explants also contained a VIP-stimulatory molecule that was immunoprecipitated by CDF/LIF antiserum, and CDF /LIF antiserum partially blocked VIP induction in explants. CDF/LIF mRNA was increased in dissociated cell cultures, in ganglion explants and in vivo after axotomy. Our results suggest that CDF/LIF released from ganglionic nonneuronal cells plays an important role in regulating VIP after axotomy. o 1994

[1] 96w Sympathetic neurons, like many other classes of neurons, normally contain one or more neuropeptides in addition to a classical or small molecule neurotransmitter. Following axotomy ofadult sympathetic neurons in the superior cervical ganglion, the content of norepinephrine, their classical transmitter, and the activity of the catecholamine synthetic enzyme, tyrosine hydroxylase, decreases (Cheah and Geffen, 1973;Hendry, 1975;Kessler and Black, 1979). These reductions in noradrenergic properties are relatively slow, taking a week or Received October I 1, 1993;accepted November 16, 1993Journal of Neurobiology, Vol. 25, No. 4, pp. 4 15-430 (1994) 0 1994 John Wiley & Sons, Inc.

[2] 166w CCC 0022-3034/94/040415-16 * To whom correspondence should be addressed more, and modest, averaging 40%-50%. In contrast, after axotomy there is a rapid and dramatic induction of at least two neuropeptides, vasoactive intestinal peptide (VIP) and substance P (SP) (Hyatt-Sachs et al., 1993;Rao et al., 1993b). At 48 h after the two major postganglionic trunks are severed, VIP levels are increased 22-fold (Hyatt-Sachs et al., 1993), and SP levels are increased 12-fold (Rao et al., 1993b). The increases in peptide content are accompanied by increases in the respective mRNAs and the appearance of VIP and SP-immunoreactive (IR) neurons and processes in the ganglion (Hyatt-Sachs et al., 1993;Rao et al., 1993b). Although there is evidence to support the notion that the decreased expression of catecholaminergic properties following axotomy results from the loss of nerve growth factor that is normally transported from terminal arbors in target tissues ( Hendry, 1975; Kessler and Black, 1979; Thoenen and Barde, 1980), it is not known what causes the striking induction of neuropeptides.

[3] 208w VIP levels also rise dramatically when either neonatal or adult ganglia are placed in explant culture (Zigmond et al., 1992;Sun et al., 1992). Within 24 h, there is a six-to tenfold increase in VIP and by 48 h, the VIP content is elevated 30-fold (Sun et al., 1992;Zigmond et al., 1992). As in the case of axotomy, the increased VIP content is accompanied by an increase in mRNA and the appearance of immunoreactive principal neurons (Zigmond et al., 1992). Explantation of ganglia into culture entails severing not only postganglionic axons but also interrupting preganglionic innervation which leads to a lack of depolarization of postganglionic neurons. Two lines of evidence indicate, however, that the VIP induction observed following explantation is primarily the consequence of axotomy and not deafferentation. First, the VIP content of the superior cervical ganglion is only modestly increased (twofold) when the preganglionic innervation is cut and the ganglion is left in situ (Hyatt-Sachs et al., 1993). Second, when sympathetic ganglion explants are depolarized, the VIP content is further increased, rather than decreased (Sun et al., 1992). VIP levels also increase when neurons from the neonatal rat SCG are dissociated and grown in culture (Nawa and Sah, 1990;Nawa and Patterson, 1990;Rao et al., 1992b;Sun et al., 1992).

[4] 25w Since VIP is induced in these two in vitro systems, they provide an opportunity to elucidate the mechanisms responsible for increasing VIP expression after axotomy.

[5] 271w Two molecules, cholinergic differentiation factor (also known as leukemia inhibitory factor; CDF/LIF) and ciliary neurotrophic factor (CNTF), that increase the expression of VIP have been identified in studies of neuropeptide expression in cultures of neurons dissociated from the superior cervical ganglia of newborn rats (Ernsberger et al., 1989;Nawa and Patterson, 1990;Nawa and Sah, 1990;Nawa et al., 1991a;Rao et al., 1992b) Although most attention has been directed toward the ability of CDF/LIF and CNTF to induce cholinergic function and reduce noradrenergic (Fukada, 1985;Saadat et al., 1989;Yamamori et al., 1989;Nawa et al., 1991b;Patterson, 1992), these two cytokines also influence the neuropeptide phenotype of cultured sympathetic neurons by inducing VIP, SP, and somatostatin and decreasing neuropeptide Y expression ( Ernsberger et al., 1989;Nawa and Patterson, 1990;Freidin and Kessler, 199 1;Nawa et al., 199 la,b;Rao et al., 1992b;Shadiack et al., 1993). It is not surprising that these iwo cytokines have such similar, if not identical actions, on sympathetic neurotransmitter properties since they share two receptor subunits and have common signalling pathways (Ip et al., 1992(Ip et al., , 1993;;Davis et al., 1993). CDF/LIF and CNTF increase VIP mRNA levels and cause VIP peptide content to rise in as little as 24 h, consistent with the possibility that one or both mediate the changes observed after axotomy ( Ernsberger et al., 1989; Nawa et al., 1991a,b; Lewis et al., 1994). In addition, CDF/LIF and CNTF are localized in or secreted from the nonneuronal cells, glial satellite, or Schwann cells and fibroblasts, present in sympathetic ganglia in situ and in vitro (Patterson and Chun, 1977;Lubbert et al., 1991;Dobrea et al., 1992;Friedman et al., 1992;Masu et al., 1993).

[6] 47w We have taken advantage of dissociated cell and explant culture systems to examine the possible roles of the differentiation factors, CDF/LIF and CNTF, in the induction of VIP that occurs after axotomy. Our findings provide evidence that CDF/LIF, but not CNTF, mediates this neuronal response to axotomy.

[1] 234w VIP is induced when sympathetic ganglia from newborn rats are dissociated and grown in cell culture. Immediately after dissociation, little VIP is detectable by radioimmunoassay. After 48 h, however, there is approximately a 20-fold increase in VIP content [Fig. 1 (a)]. Since the cell suspension obtained ii-om the dissociated ganglia contained nonneuronal cells and the cultures were grown without antimitotic agents, nonneuronal cells were present. To determine whether the presence of nonneuronal cells contributed to the increase in VIP expression, cultures were established that were enriched in neurons and depleted of nonneuronal cells. This was accomplished by incubating the cell suspension obtained after enzymatic dissociation of the ganglia in tissue culture dishes for 3 h; during this period many nonneuronal cells attached to the treated plastic, while neurons did not. The medium containing the nonadherent cells, predominantly neurons, was then plated into 96-well plates. The preplating resulted in a substantial decrease in the number of nonneuronal cells [Fig. 1 (b,c)] . The VIP content of the neuron-enriched cultures was significantly less than that of the mixed neuron and nonneuronal cell cultures: although the neuron-enriched wells contained approximately twice as many neurons as the mixed wells, they contained less than half as much VIP [Fig. 1 (a)]. When the VIP content of the neuronenriched cultures is calculated on a pg/neuron basis, it was approximately 1 5 8 that of the mixed cultures after 48 h.

[2] 43w To determine whether the effect of the nonneuronal cells on VIP expression was mediated by a diffusible factor, we tested the effects of medium conditioned by nonneuronal cells. Sympathetic ganglia were dissociated and cultures of ganglionic nonneuronal cells were obtained by preplating the

[3] 11w . . * I 30 -, a t=O t=48 t=3 k48

[4] 80w cell suspension. The adherent nonneuronal cells were grown in the absence of NGF. Medium conditioned by the ganglionic nonneuronal cell cultures (GNCM ) was used to treat neuron-enriched cultures. After 48 h, there was approximately fourfold more VIP-IR detected by radioimmunoassay in the cultures grown in the presence of nonneuronal cell-conditioned medium than in its absence (Fig. 2 ) . Thus, medium conditioned by ganglionic nonneuronal cells contained a VIP-inducing activity and could substitute for the presence of nonneuronal cells.

[5] 550w Two major classes of nonneuronal cells are present in neonatal sympathetic ganglia, fibroblasts, and glial cells, both satellite and Schwann cells (Hall and Landis, 1992). To determine whether one or both of these classes were present in the nonneuronal cell cultures, we characterized the nonneuronal cells using two immunological probes: an antiserum raised against fibronectin, an extracellular matrix component, which identifies fibroblasts and the monoclonal antibody 2 17c which recognizes p75, the low-affinity NGF receptor, and identifies satellite and Schwann cells (Kumar et al., 1991). The nonneuronal cell cultures contained both classes of cells. with at least two- The remaining cells were placed in a 35-mm tissue culture dish in complete L15C0, medium. After 3-4 h, during which time many of the nonneuronal cells attached, the nonadherent cells were col-lected, centrifuged, and resuspended in F12-defined medium. This population of cells, enriched in neurons, was either assayed for their content ofVIP immediately( neuron-enriched, t = 3, hatched bar) or grown in 96-well plates in F12-defined medium for another 45 h (neuronenriched, t = 48, hatched bar). VIP was determined by radioimmunoassay. The t = 0 and t = 3 determinations represent the mean oftwo wells each. The other determinations represent the mean & S.E.M. of eight wells each. The cultures of neurons and nonneuronal cells contained an average of 1144 ? 56 neurons/well, whereas the neuron-enriched cultures had an average of 2364 ? 7 1 neurons/well. In the histogram, the VIP is expressed as pg/well. When calculated on a pg/neuron basis, the VIP content ofthe neuron-enriched cultures was approximately 15% of that of the mixed cultures. The values were compared using a one-way ANOVA and a post hoc t test. *f = 48 compared with t = 0, p < 0.001; **t = 48 compared with t = 3,p < 0,001; ***t = 48 (mixed neurons and nonneuronal cells) is compared with t = 48 ( neuron-enriched), p < 0.00 1. (b) and (c) The morphological appearance of mixed (b) and neuron-enriched (c) cultures is compared after 48 h. Note the large reduction in the number of nonneuronal cells in cultures enriched for neurons by preplating. In both conditions, the neurons appear phase-bright and have extended neurites. Arrows indicate nonneuronal cells. The focus is on the nonneuronal cells in ( b ) and on the neurons in (c). F12-defined medium was conditioned for 48 h by ganglionic nonneuronal cells from neonatal superior cervical ganglia. The nonneuronal cell-conditioned medium (GNCM) and control F12-defined medium (control) were concentrated in Centriprep tubes with a 10 kDa cutoff. The concentrated media were diluted with fresh F12-defined medium to which N G F had been added so that final concentration of the conditioned medium was approximately 1X and were used to treat neuronenriched cultures. After 48 h, the neurons in each well were counted, peptides were extracted, and VIP-IR was determined by radioimmunoassay. On average, 101 4 f 52 neurons were present in each well grown with F12defined medium and 991 i 51 neurons in each well grown with nonneuronal cell conditioned medium. Each bar represents the mean of six culture wells. The VIP content in the wells treated with nonneuronal cell conditioned medium is significantly different ( p < 0.0001) from that in the wells treated with control medium using a two-tailed Student's t test.

[6] 10w thirds possessing immunoreactivity for 2 17c (Fig. 3 ) .

[7] 251w Two molecules that induce VIP in cultured sympathetic neurons, CDF/LIF and CNTF, have been purified and cloned (Lin et al., 1989;Stockli et al., 1989;Yamamon et al., 1989;Nawa and Patterson, 1990;Nawaetal., 1991a;Raoetal., 1992b).CDF/ LIF is secreted by fibroblasts and may also be produced by Schwann and satellite cells (Patterson and Chun, 1977; Lubbert et al., 199 1 ; Shadiack et al., 1993). CNTF is present in high concentrations in myelinating Schwann cells and can be released into the medium by Schwann cells cultured from sciatic nerve by an undefined mechanism (Stockli et al., 1991;Dobrea et al., 1992;Friedman et al., 1992;Meyer et al., 1992). To determine whether either differentiation factor was present in the ganglionic nonneuronal cell-conditioned medium, we performed immunoprecipitation experiments with antisera that recognized either rat CDF/LIF or CNTF (Fig. 4). Preincubation with anti-CDF/LIF completely abolished the response to the conditioned medium observed in neuron-enriched cultures. Antisera to CDF/LIF also appeared to potentiate the effect of CNTF on VIP induction. In contrast, anti-CNTF antiserum had no detectable effect on the induction of VIP by nonneuronal cellconditioned medium even though it reduced the response to exogenous CNTF. Since the immunoprecipitation experiments failed to provide evidence for the presence of CNTF in medium-conditioned by ganglionic nonneuronal cells, we used survival of E8 chick ciliary neurons as a bioassay for CNTF (Barbin et al., 1984;Manthorpe et al., 1986). While exogenous CNTF (1 ng/ml) supported the survival of and process outgrowth from ciliary neurons, medium conditioned by ganglionic nonneuronal cells did not (data not shown).

[8] 108w The failure of the GNCM to promote ciliary neuron survival is not due to toxicity since the addition of CNTF ( 5 ng/ml) resulted in survival even in the presence of GNCM. The results of the immunoprecipitation studies suggested that ganglionic nonneuronal cells released CDF/LIF. To examine the expression of CDF/LIF by nonneuronal cells further, we performed a semiquantitative analysis of the presence of mRNA encoding CDF/LlF using reverse transcriptase-polymerase chain reaction (RT-PCR ) . After 48 h in culture, CDF/LIF mRNA levels in ganglionic nonneuronal cells were greater than that in freshly dissected ganglia when levels were normalized against those of P-actin message (Fig. 5 ) .

[9] 86w When superior cervical ganglia of adult rats are placed in organ culture, VIP content is increased 30-fold by 48 h (Zigmond et al., 1992). To determine whether a soluble factor released by cells in the ganglion was responsible, we collected medium conditioned by explanted superior cervical ganglia and tested its ability to induce VIP in neuronenriched cultures from neonatal superior cervical ganglia. Medium conditioned by explanted ganglia, like that conditioned by ganglionic nonneuronal cells in dissociated culture, caused a fourfold increase in VIP content (Fig. 6).

[10] 171w To characterize the VIP-inducing factor in medium conditioned by explanted ganglia, we performed immunoprecipitation experiments with the rat CDF/LIF antisera. The CDF/LIF antiserum had no detectable effect on the induction of VIP by exogenous CNTF but significantly reduced the induction by ganglion-conditioned medium [Fig. 7 (a)], indicating that one of the components mediating the VIP increase in medium conditioned by explanted ganglia is CDF/LIF. To determine if the presence of CDF/LIF in the ganglion conditioned medium was associated with the presence of CDF/ LIF mRNA, we utilized semiquantitative RT-PCR to detect CDF/LIF message in explant cultures. CDF/ LIF message was detectable in ganglia immediately after dissection but levels rose dramatically within 48 h (Fig. 7). In contrast, no change in 0-actin mRNA was evident. When the content of CDF/LIF mRNA in control and cultured ganglia was normalized against the amount of /3-actin mRNA, a large increase in CDF/LIF message was evident. These findings provided evidence that CDF/LIF was made in the explants and released into the medium in biologically active amounts.

[11] 237w To determine whether CDF/LIF was necessary for the induction of VIP in explanted ganglia, we cultured ganglia in medium containing the antiserum to rat CDF/LIF which is neutralizing as well as precipitating (Fukada and Towle, 1992). As can be seen in Figure 8(a), the CDF/LIF antiserum, added at a final concentration of 8%, suppressed the VIP induction normally seen in explanted ganglia by more than 60%. In contrast, an equivalent concentration of nonimmune rabbit serum had no effect on VIP induction. Two experiments were undertaken to ascertain whether the failure to detect an increase in VIP levels in the presence of the neutralizing CDF/ LIF antiserum was due to toxicity. First, we compared in 1 -pm plastic sections the morphological appearance of neurons in ganglia grown with or without neutralizing antisera. We found no evidence for degenerative changes, such as pyknotic nuclei or vacuoles (Sun et al., 1992) and the neurons in ganglia cultured in normal rabbit serum were indistinguishable from those in ganglia incubated with the neutralizing antiserum [Fig. 8 (b,c)] . Second, when recombinant human CDF/LIF, which is not blocked by the antiserum raised against rat CDF/LIF (Fukada and Towle, 1992;Rao et al., 1992a), was added to the explanted ganglia, ganglionic levels of VIP increased even in the presence of the antiserum. Thus, in the presence of the blocking antiserum the neurons are morphologically normal and able to respond to exogenous factors by elevating VIP.

[12] 22w 0 no ab anti CNTF anti LIF 70 G N C M 40 no ab anti CNTF anti LIF 70 1 ...

[13] 90w The induction of VIP observed when ganglia are placed in explant culture mimics that observed in vivo when postganglionic axons in the internal and external carotid nerve are cut (Sun et al., 1992;Zigmond et al., 1992;Hyatt-Sachs et al., 1993). A detailed comparison of the changes in VIP content following explantation and axotomy revealed that the rate of increase and its magnitude were very similar [Fig. 9 (a,b)] . These similarities raised the possibility that CDF/ LIF contributed to the induction not only in vitro but also in vivo after axotomy.

[14] 21w We therefore determined whether CDF/LIF mRNA levels were altered after the postganglionic trunks were transected and the ganglia left in situ.

[15] 101w Semiquantitative RT-PCR analysis showed that CDF/ LIF mRNA levels are significantly increased 48 h after axotomy in comparison to those of sham-operated animals (Fig. 10). quantitative RT-PCR are increased in cultures of ganglionic nonneuronal cells and in explanted ganglia. When the induction of VIP in sympathetic ganglia after explantation is compared with that after axotomy, both the magnitude and the time of induction are similar. In addition, we find that CDF/LIF mRNA is increased in sympathetic ganglia after axotomy. Taken together, our data provide evidence that CDF/LIF plays a role in the induction of VIP both in vitro and in vivo.

[1] 159w Several lines of evidence suggest that while CNTF can induce VIP expression in dissociated sympathetic neuron cultures (Ernsberger et al., 1989;Lewis et al., 1994;Rao et al., 1992b), it does not normally contribute to VIP induction in dissociated cultures or in explants. GNCM, which contains VIP-inducing activity does not promote the survival of ciliary neurons. Since the half-maximal response of CNTF for ciliary survival is tenfold lower than that for VIP induction, ciliary neuron survival is a very sensitive assay for the presence or absence of CNTF (Stockli et al., 1989;Leung et al., 1992;Rao et al., 1992b). Further, antibodies to CNTF previously shown to immunoprecipitate CNTF from complex mixtures (Rao et al., 1992a) do not immunoprecipitate the VIP-inducing activity in GNCM. The observation that nonneuronal cells in sympathetic ganglia, in contrast to those in sensory ganglia, do not contain significant levels of CNTF mRNA or protein in vivo is consistent with these results (Dobrea et al., 1992;Friedman et al., 1992).

[2] 89w In contrast to the lack of evidence supporting a role for CNTF, our findings indicate that CDF/ LIF plays a role in regulating VIP expression in dissociated and explant cultures. First, CDF/ LIF mRNA is elevated in both dissociated and explant cultures. Second, the increase in CDF/LIF mRNA is associated with the presence of VIP-inducing activity in medium conditioned by ganglionic nonneuronal cells or explanted ganglia that is immunoprecipitated with CDF/ LIF antibodies. Third, antibody blocking experiments provide evidence that CDFILIF is necessary for VIP induction in explanted ganglia.

[3] 72w We found that neutralizing antibodies specific for rat CDF/LIF (Fukada and Towle, 1992) partially block the increase in VIP in explanted ganglia. The suppression of VIP induction by the polyclonal antiserum is not due to toxicity. There is no evidence of neuronal degeneration in 1 -pm plastic sections (see Sun et al., 1992, for comparison). Further, we observed a significant increase in VIP expression in response to human recombinant CDF/ ' "1

[4] 158w Medium conditioned by superior cervical ganglion explant cultures from adult rats induces VIP in neuron-enriched cultures. Fl2-defined medium was incubated for 48 h at 37°C in the presence (SCM) or absence (control ) of adult superior cervical ganglia explanted from adult rats. Two ganglia were used to condition 750 pl medium. Both media were concentrated in Centriprep tubes with a 10 kDa cutoff. The concentrated media were diluted with fresh F12-defined medium containing NGF so that the final concentration of the conditioned medium was approximately 1X. This medium was used to treat neuron-enriched cultures of neonatal superior cervical ganglia. After 48 h, the peptides were extracted from the culture wells and VIP-IR was determined by radioimmunoassay. Each bar represents the mean of four culture wells. The VIP content of the wells treated with conditioned medium was significantly different ( p < 0.0001) from that of the wells treated with F12 defined medium using a two-tailed Student's t test.

[5] 39w cmtrd 50 1 no ab anti LIF 40 5 8 30 LI) n a 20 Y L > 10 0 no ab anti LIF ' 30-LI) 0 . t i ! 20v E l i 5 * 10-0 .

[6] 270w Immunoprecipitation of ganglion conditioned medium with an antiserum to rat CDFILIF decreases the induction of VIP. F12-defined medium and medium conditioned by sympathetic ganglion explants were concentrated tenfold. The concentrated conditioned medium (SCM), concentrated F12-defined medium to which 10 ng/ml of recombinant rat CNTF (CNTF) concentrated F12-defined medium (control), were incubated either without added antiserum (no ab) or with antiserum generated against purified rat CDF/LIF (anti-LIF) in 200 pl of reaction mixture at 4°C overnight. The antisera and bound ligands were precipitated with protein A-sepharose beads. After immunoprecipitation, the conditioned medium and the F12 media were each added to 2 ml of fresh F12-defined medium containing NGF and used to treat neuron-enriched cultures from neonatal superior cervical ganglia. After 48 h, the peptides were extracted, and VIP-IR was determined by radioimmunoassay. Each bar represents the mean of 16 culture wells. When the experimental values were compared with control (no ab) within each panel using a two-way ANOVA plus Tukey's protected t test, thep value is as follows: * p < 0.0 1. In the lower right panel is shown the results of RT-PCR analysis of CDF/LIF expression in control and explanted ganglia. Total RNA was extracted from either adult superior cervical ganglia immediately after dissection ( t = 0) or ganglia after 48 h in culture in F12 medium ( t = 48). About 0.5 fig of total RNA was used for RT-PCR amplification. In addition to CDF/LIF DNA, P-actin cDNA was amplified as an internal control for both transcription and amplification. PCR products were checked by Southern blot analysis with oligonucleotide probes recognizing internal sequences in the amplified products.

[7] 142w LIF in the presence of the antiserum generated against rat CDF/LIF. Human CDF/LIF, which is not recognized by the antiserum (Fukada and Towle, 1992; Rao et al., 1992a), presumably acts via the same receptor as does endogenous CDF/ LIF. Thus, the neurons remain functionally responsive in the presence of antiserum. Although, in principle, the antiserum we used could have cross-reacted with other molecules immunologically related to, but distinct from, CDF/LIF, we .. believe this is unlikely. The antiserum raised against purified rat CDF/ LIF does not recognize either recombinant human CDF/LIF, which is greater than 80% homologous to rat CDF/LIF (Yamamon et al., 1989), or recombinant rat CNTF which shares many neuronal activities, and a common topological structure with CDF/LIF (Bazan, 1991;Patterson, 1992;Rao and Landis, 1993). In addition, the RT-PCR studies provide evidence for the presence of CDF/LIF mRNA in explanted ganglia.

[8] 150w Although the CDF/LIF antiserum suppresses the induction of VIP in explanted ganglia, the suppression is only partial. There are several possible explanations for this finding. First, effective concentrations of CDF/LIF antibodies may not be present throughout the ganglia. Alternatively or additionally, since increases in CDF/ LIF mRNA occur very rapidly in response to injury (Minami et al.,199 1 ), and increases in VIP mRNA and protein can occur very rapidly in sympathetic neurons and related cell lines in response to CDF/ LIF and CNTF (Lewis et al., 1994;Symes et al., 1993), VIP induction may have already been initiated before the antibody blockade becomes effective. Finally, molecules other than CDF/ LIF could participate in the induction of VIP. Since VIP induction is virtually absent when sympathetic ganglia of transgenic mice that lack CDF/LIF are placed in explant culture (Rao et al., 1993a), however, it seems unlikely that this is the case.

[9] 168w Our results indicate that CDF/ LIF is produced by ganglionic nonneuronal cells. Medium conditioned by cultures of nonneuronal cells dissociated from sympathetic ganglia contains VIP-inducing activity that is immunoprecipitable by CDF/ LIF antibodies. CDF/LIF mRNA is detectable in nonneuronal cell cultures and levels are higher after 48 h than in freshly dissected ganglia. The ganglionic nonneuronal cell cultures, which are obtained from neonatal sympathetic ganglia by selecting adherent cells and growing them without NGF, contain fibroblasts and glial satellite and Schwann Ganglia dissected from adult rats were grown in explant culture in F12-defined medium containing either 8% serum from a normal rabbit or 8% serum from a rabbit immunized with rat CDF/ LIF purified from heart cellconditioned medium. After 48 h, the ganglia were fixed and embedded in plastic. Semithin sections were cut and stained with toluidine blue. Neurons grown in either 8% rabbit serum (b) or 8% anti LIF antiserum ( c ) appear healthy with large clear nuclei, prominent nucleoli and even, granular cytoplasm ( 500X

[10] 94w ). cells. CDF/LIF is secreted by fibroblasts and may also be produced by Schwann and satellite cells (Patterson and Chun, 1977; Lubbert et al., 1991; Shadiack et al., 1993). Either or both cell classes could be responsible for the VIP induction in vitro. At present, we cannot rule out the possibility that neurons also make CDF/LIF. Although neuronenriched cultures contain significantly less VIP than do mixed neuron/ nonneuronal cell cultures, the VIP content is elevated. Since the neuronenriched cultures still contain nonneuronal cells, these residual cells could be responsible for the increase in VIP.

[11] 202w In addition to VIP, CDF/LIF induces cholinergic function, SP, and somatostatin in cultures of sympathetic neurons dissociated from neonatal ganglia (Yamamori et al., 1989; Nawa and Patterson, 1990; Freidin and Kessler, 1991; Nawa et al.. 1991a,b; Lewis et al., 1994; Shadiack et al., 1993). Since our results indicate that ganglionic nonneuronal cells can synthesize and release CDF/LIF, it seems likely that CDF/LIF is responsible for the induction of cholinergic function previously reported in mixed cultures of sympathetic neurons and ganglionic nonneuronal cells and in explant cultures (O'Lague et al., 1974, 1978; Patterson and Chun, 1974; Johnson et al., 1980). When neonatal or adult ganglia are cultured as explants, the expression of SP and somatostatin is increased (Kessleret al., 1981, 1983; Roach et al., 1987; Sun et al., 1992 1 Two agents, dexamethasone and interleukin 10 (IL-I@), which inhibit and elevate the increase in SP, respectively (Kessler et al., 1983: Freidin and Kessler, 199 1 ; Hart et al., 199 1, Shadiack et al., 1993), have similar effects on VIP levels in explant cultures (Rao et al., 1993b). Given the similarities in the regulation of SP and VIP, CDF/LIF is likely to regulate not only VIP but also SP expression in explant cultures.

[12] 325w Our results suggest that following the transection of axons of sympathetic neurons, either as a consequence of explantation into organ culture or axotomy in situ, ganglionic nonneuronal cells increase their expression and secretion of CDF/ LIF. We found that CDF/LIF mRNA assayed by semiquantitative RT-PCR increases after explanation (see also Shadiack et al., 1993) and after axotomy in situ, raising the question ofwhether the signal( s) immediately responsible for inducing CDF/ LIF in nonneuronal cells after neuronal injury are generated within the ganglion. Examination of the regulation of SP expression in culture suggests that IL-1 P is involved (Jonakait et al., 1990;Freidin and Kessler, 1991;Hart et al., 1991;Freidin et al., 1992;Shadiack et al., 1993). Cultured sympathetic neurons synthesize and release IL-10 (Freidin et al., 1992). IL-1 / 3 increases SP expression in sympathetic neurons in dissociated cell and explant cultures (Jonakait and Schotland, 1990;Freidin and Kessler, 1991;Hart et al., 1991, Shadiack et al., 1993). The induction of SP that normally occurs in mixed cultures of neurons and nonneuronal cells is significantly reduced by an IL-1 receptor antagonist (Freidin et al., 1992). Finally, the SP induction by IL-1P requires the presence of nonneuronal cells and therefore presumably involves an intermediate molecule released by nonneuronal cells (Freidin and Kessler, 1991;Freidin et al., 1992;Shadiacket al., 1993). CDF/LIFisan excellent candidate for this intermediate molecule (Shadiack et d., 1993) since it induces neuropeptide expression, is secreted by ganglionic nonneuronal cells, and is induced by IL-10 in cells of mesenchy-ma1 origin (Lubbert et al., 1991;Wetzler ei al., 1991;Hamilton et al., 1993) and in sympathetic ganglia in explant cultures (Shadiack et al., 1993). IL-lP also increases VIP expression in SCG explant cultures, though the magnitude of the change in VIP is less than that of SP (Rao et al., 1993b). While the studies summarized above provide strong evidence for the role of IL-10 in peptide induction in culture, evidence is presently lacking for its participation in the response to axotomy in vivo.

[13] 309w Several lines of evidence indicate that the VIP induction that occurs in explanted ganglia mimics the VIP induction observed after axotomy in situ, consistent with the possibility that CDF/ LIF is a mediator in both inductions. Both experimental paradigms share a common means of neuronal injury, axon transection. In both cases, there is a rapid and large elevation of VIP mRNA and protein (Zigmond et al., 1992;Hyatt-Sachs et al., 1993). The magnitude and the time course of the increases in VIP in explanted ganglia and after axotomy in situ are similar. Further, using semiquantitative RT-PCR analysis, we detect a significant elevation of CDF/LIF mRNA in both explanted and axotomized ganglia, indicating that in vivo as in explant cultures, the increase in VIP expression is correlated with an increase in CDF/LIF expression. By analogy with the requirement for CDF/ LIF for the increase in VIP seen in explant culture, CDF/LIF is likely to contribute to the elevation of VIP levels observed after axotomy. Consistent with this hypothesis, there is significant suppression of VIP induction after axotomy in mutant mice lacking CDF/LIF (Rao et al., 1993a). When ganglia from CDF/LIF-deficient mice are placed in explant culture, little increase in VIP content or the number of immunoreactive neurons is detected. Further, when axotomy is performed in situ, the increase in VIP content is suppressed by 60%-70% and while the number of VIP-immunoreactive neurons is increased, the increase is much less than that observed in normal mice. It is also of interest that the increase of tachykinin-immunoreactive neurons seen after axotomy in wild-type ganglia is also suppressed. Taken together, although our results do not exclude contributions of other peptideinducing factors (Nawa and Patterson, 1990;Nawa and Sah, 1990;Coulombe and Nishi, I99 1 ) to the regulation of VIP expression after axotomy in vivo, they do provide strong evidence that CDF/ LIF plays a role.

[1] 277w Cell culture reagents were obtained from Gibco (Grand Island, NY) and from Collaborative Biomedical Becton Dickinson Labware (Bedford, MA). Culture plates were obtained from Corning (Corning, NY). Nerve growth factor (NGF) was purchased from AUSTRAL Biologicals (San Ramon, CA). Dispase was obtained from Boehringer-Mannheim (Indianapolis, IN) and collagenase from Worthington Biochemicals (Freehold, NJ). For the VIP radioimmunoassay, the rabbit anti-VIP, the goat anti-rabbit IgG, and the normal rabbit serum were purchased from Peninsula Labs (Belmont, CA) and the ['251]-VIP from NEN (Boston, MA). Centricon filters were purchased from Amicon (Danvers, MA). The reverse transcription polymerase chain reaction reagents were obtained as a kit from Perkin Elmer Cetus (Norwalk, CT). [ 32P] r-ATP was purchased from ICN Radiochemicals (Costa Mesa, CA). The LIF and p-actin oligonucleotides were purchased from Oligos Etc, Inc (Wilsonville, OR). The 217c monoclonal antibody was the kind gift ofDr. Jean deVellis (UCLA, CA). Rabbit antifibronectiri was obtained from Telios Pharmaceuticals (San Diego, CA), fluorescein-conjugated goat anti-rabbit from Cappel Oraganon Teknika (Westchester, PA), and biotin-conjugated goat anti-rabbit, Texas red-conjugated to avidin and Texas red-conjugated goat antimouse from Jackson Research Labs (West Grove, PA). The recombinant rat CNTF and the antiserum gener-ated against it was the kind gift of Dr. Donna Momssey at Regeneron Pharmaceuticals (Tarrytown, NY). Human recombinant CDF/LIF was purchased from R and D Systems, Inc. (Minneapolis, MN). An antiserum generated against CDF/LIF purified from heart cell-conditioned medium was used both in immunoprecipitation and in blocking experiments and was the kind gift of Dr. Keiko Fukada (Fukada et al., 199 I;Fukada and Towle, 1992; SUNY Downstate, NY). Animals were obtained from Zivic Miller (Zelienople, PA). Other chemicals were purchased from Sigma (St. Louis, MO).

[2] 92w The levels of VIP-like immunoreactivity were determined by radioimmunoassay as previously described (Zigmond et al., 1992;Sun et al., 1992;Hyatt-Sachs et al., 1993). These data are referred to in the text as levels of VIP. In brief, cultures and ganglia were homogenized in 2Nacetic acid. After boiling the cultures for 5 rnin and the ganglia for 20 min, samples were centrifuged for 1 rnin in an Eppendorf microfuge. The supernatants were lyophilized and stored at -70°C for subsequent assays. The primary antiserum has been previously shown to display minimal cross-reactivity with other peptides.

[3] 114w To characterize the nonneuronal cells present in the cultures, nonneuronal cells were grown on glass coverslips. After 48 h, the cells were fixed with 4% paraformaldehyde in 0.1 Mphosphate buffer at room temperature for 20 rnin and then washed with PBS. The coverslips were incubated with the primary antibodies, 2 17c ( 1 : 10 monoclonal supernatant) and anti-fibronectin ( I : 100) in PBS containing 20% goat serum, for 30 rnin at room temperature, washed with PBS, and incubated with the secondary antisera, Texas Red-conjugated goat anti-mouse and fluorescein-conjugated goat anti-rabbit, for 30 rnin at room temperature ( 1:200 dilution). The coverslips were washed and then mounted in PBS:glycerol containing 2% n-propyl gallate.

[4] 142w lmmunoprecipitation Experiments with CDF/LIF and CNTF Antibodies 100 pl of tenfold-concentrated medium conditioned by either nonneuronal cells or explanted sympathetic ganglia were added to buffer (PBS. pH 7.3, with 2% bovine serum albumin, 0.2% Triton X-100, and 0.02% polyethylene glycol 6000 and 1 12). 40 pl of blocking antiserum generated against CDF/ LIF purified from rat heart cellconditioned medium (Fukada and Towle, 1992) or of an antiserum generated against recombinant rat CNTF were added to each vial. After an overnight incubation, the antigen-antibody complex was absorbed to 40 pl of Protein A Sepharose for an additional 2 h at room temperature. The bound complexes were separated by centrifugation. Cultures were treated with the supernatant diluted into F12 medium supplemented with 50 ng/ml NGF. We used 100 p1 of conditioned medium incubated without the antibody and treated as described above as a control.

[5] 71w A GeneAmp RNA PCR kit (Perkin-Elmer Cetus, Norwalk, CT) was used according to the protocol provided to detect LIF message in nonneuronal cell cultures, in explant cultures, and in the superior cervical ganglion after axotomy in situ. About 0.5 pg total RNA and random hexamers were used to perform reverse transcription in a DNA thermal cycler at 42°C for 45 min, then 99°C for 5 min, and 5°C for 5 min.

[6] 215w For PCR amplification, an upstream primer was synthesized from the rat CDFILIF sequence from code 179-199 (5'-ATGCCCTCTTTATTTCCTATT-3') and a downstream primer from code: 589-569 (5'-CACTTA-TGACTTGCTTGTATG-3'). 0-Actin cDNA was amplified as an internal control using sequences from the human gene. The upstream (3-actin primer was from and the downstream primer from 1498-1479 (5'-TGT-AGAAGGTGTGGTGCCAG-3'). The PCR reaction was performed in a DNA thermal cycler programmed for four sequences as follows: ( 1 ) Step-cycle: 94°C. 5 min followed by 53"C, 5 min for I cycle. ( 2 ) Step-cycle: 72"C, 1.5 min followed by 94"C, 45 s followed by 53"C, 45 s for 30 cycles. (3) Step-cycle: 72"C, 10 niin for 1 cycle and (4) Soak: 4°C. The PCR products were first checked by size on a 1.5% agarose gel by ethidium bromide staining and by comparison with DNA size 1357-1 376 (5'-GTGGGCATGGGTCAGAAGGA-3) markers. Identification of the PCR products was verified by Southern blot analysis with end labelled DNA oligonucleotides synthesized from sequences present between the two sets of pnmers. For CDF/LlF, the probe was from 485-504 ( 5'-GCCATGTGGATGTGCCCTGT-3') and for actin from 139 1 -14 10 ( 5 '-ACGAGGCCCAGA-GCAAGAGA-3'). This sequence is identical in the human and rat 0-actin genes. QuickHyb solution from Stratagene was used in the Southern blot analysis according to the manufacturer's directions.

[1] 105w Cultures of rat sympathetic neurons were prepared as described by Hawrot and Patterson ( 1979). Neurons from superior cervical ganglia were dissociated enzymatically with dispase ( 5 mg/ml) and collagenase ( I mg/ ml). Dissociated neurons, 1000-2000 per well, were plated in 94-well plates sequentially coated with poly-llysine (0.1 mg/ml) and laminin (5-10 pglml). Cells were grown in FI2-defined medium ( 5 pg/ml insulin, 20 n M progesterone, 30 p M selenium, 5 mg/ml bovine serum albumin, 100 pg/ml transferrin, 100 pit4 putrescine) supplemented with NGF (50 ng/ml), penicillin (50 units/ml), and streptomycin (50 pg/ml). Peptide was extracted and assayed after 48 h in culture.

[2] 86w To obtain neuron-enriched cultures, cells were dissociated as described above and then preplated for 3-4 h on uncoated 35-mm tissue culture plastic dishes in 1,-15 C 0 2 medium containing NGF, penicillin ( 100 units/ ml), streptomycin ( 100 pg/ml) and 5% rat serum. During this period, many nonneuronal cells attached to the substrate. The nonadherent cells which were primarily neurons, were collected, resuspended in F I2 defined medium and plated in 96-well dishes. Peptide was extracted and assayed after an additional 45 h in culture.

[3] 101w To obtain ganglionic nonneuronal cells, we preplated a suspension of dissociated cells for 3-4 h on uncoated 35-mm tissue culture plastic dishes in L-15 CO, medium containing NGF, penicillin ( 100 units/ml), streptomycin (100 pg/ml), and 5% rat serum. The nonadherent neurons were removed by washing, and the adherent cells were grown in F12-defined medium without NGF. Each dish contained nonneuronal cells obtained from 15 ganglia. After 48 h of culture, medium conditioned by the nonneuronal cells was collected. This medium, after concentration and redilution with fresh FI 2 medium with 50 ng/ml NGF added, was used to treat neuronal cultures.

[4] 89w To obtain explant cultures, we dissected superior cervical ganglia from Sprague-Dawley rats 4-5 weeks of age. The ganglia were placed in organ culture on Millipore filters on stainless-steel rafts. Details of the procedure have been described previously (Zigmond and MacKay, 1974;Zigmond et al., 1992) except that F12defined medium was used. After 48 h of culture of two adult ganglia in 750 jd medium, the medium was col-lected, concentrated, and rediluted in fresh F12 medium and used to treat dissociated neuron cultures after the addition of 50 ng/ml of NGF.

[5] 130w Explant cultures were treated with a blocking antiserum raised against CDF/LIF purified from medium conditioned by rat heart cells. The blocking antiserum or normal rabbit antiserum were used diluted 1 : 12 in F12defined medium. In some experiments, antiserumtreated explants were also treated with recombinant human LIF ( 100 ng/ml), which is not recognized by the blocking antiserum (Fukada and Towle, 1992). To ascertain whether treatment with the antiserum had adverse effects on neuronal survival, explants were immersion-fixed in 2% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer, post-fixed with 2% osmium tetroxide in 0. I A 4 cacodylate buffer, dehydrated, and then embedded in Epon as previously described (Sun et al., 1992). Plastic sections ( 1 pm) were stained with toluidine blue and examined with the light microscope.

[6] 78w Ciliary ganglia were dissected from embryonic day 8 (E8) chicks, dissociated, and plated in DMEM with 10% fetal calf serum as described by Varon ( 1979). We plated 1000-2000 neurons directly into medium in 96-well culture plates coated sequentially with polylysine and laminin. The cultures were incubated for 24-48 h at 37"C, and fixed with 2% glutaraldehyde as described by Barbin et al. ( 1984). The number of surviving neurons was determined by counting phase-bright cells with neurites.