Journal of Ncurochemisfry Raven Press, Ltd., New York 0 1991 International Society for Neurochemistry
Skeletal Muscle Proteins Stimulate Cholinergic Differentiation of Human Neuroblastorna Cells *?$James L. McManaman and *?Frances G. Crawford *Wagner ALS Research Laboratory and t The Jerry Lewis Neuromuscular Disease Research Center, Department of Neurology, Division of Neuroscience, and $Program in Cellular and Molecular Biology, Baylor College of Medicine, Houston, Texas, U S A .
Abstract: Extracts of rat skeletal muscle contain substances that enhance the development of choline acetyltransferase (ChAT) in the cholinergic human neuroblastoma cell line LA-N-2. The ChAT enhancing activity in muscle extract was purified to homogeneity by preparative gel electrophoresis and reverse-phase HPLC. The active factor is biochemically and immunologically identical to ChAT development factor, (CDF), the skeletal muscle factor that enhances ChAT activity in enriched cultures of embryonic rat motoneurons and rescues motoneurons from naturally occumng cell death in vivo. CDF increases the specific ChAT activity of LA-N-2 cells fivefold after 6 days in culture, but does not affect their growth or metabolic activity. Basic fibroblast growth factor also in-
creases ChAT activity in LA-N-2 cells and its effect is additive with that of CDF. In contrast, neither insulin-like growth factor- I , epidermal growth factor, nor nerve growth factor affected the ChAT activity of LA-N-2 cells. Our study demonstrates for the first time that CDF can directly affect the development of neuronal properties in a homogeneous population of cells, and that the effects of CDF are separate from those of other types of trophic factors. Key Words: Muscle proteins-Purification-Neuronal development-cholinergic enzymes-Neurotrophic factors-Neuroblastoma cells. McManaman J. L. and Crawford F. G. Skeletal muscle proteins stimulate cholinergic differentiation of human neuroblastoma cells. J . Neurochem. 57, 258-266 (199 1).
The functional differentiation of cholinergic neurons is associated with an increase in the levels of choline acetyltransferase (ChAT), the enzyme responsible for the synthesis of acetylcholine (Burt, 1975; Pilar et al., 1981). Although the factors that regulate the expression of this enzyme are, for the most part, unknown, increasing evidence supports the involvement of exogenous target-derived factors. This evidence is based primarily on the effects of isolated neurotrophic factors and tissue extracts on the development of ChAT activity in primary cultures of embryonic neurons (Hefti
et al., 1985; McManaman et al., 1988a,b; Saadat et al., 1989; Yamamori et al., 1989). Interpretation of these studies, however, has been complicated both by the heterogeneous nature of primary cultures and by effects of target-derived factors on neuronal survival (Martinou et al., 1989; McManaman et al., 1990). Use of cholinergic cell lines provides a means of investigating the effects of neurotrophic factors on the development of cholinergic properties in a homogeneous population of cells, in which survival is not a complicating factor. Neuronal cell lines are important model systems for
Received October 9, 1990; revised manuscript received December 20. 1990; accepted December 2 1, 1990. Address correspondence and reprint requests to Dr. J. McManaman at Department of Neurology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, U.S.A. Abbreviations used: bFGF, basic fibroblast growth factor; CDF, ChAT development factor; ChAT, choline acetyltransferase; DEAE, biologically active fraction obtained following chromatography on DEAE cellulose; E, embryonic day; EGF, epidermal growth factor; GI OOA, Sephadex G 100 void volume fraction that contains a high molecular weight form of ChAT enhancing activity;GIOOB, Sephadex
G 100 fraction that contains CDF; HAP-7, biologically active fraction obtained following chromatography on hydroxylapatite at pH 7.0; HAP-9, biologically active fraction obtained following chromatography on hydroxylapatite at pH 9.0; IGF-1, insulin-like growth factorI ; IgG, immunoglobulin G; LGG-5, identical to LGG-15, but containing 5% fetal bovine serum; LGG-15, Leibovitz's L-15 medium containing 2 mM glutamine, 0.05 mg/ml gentamicin, and supplemented with 15% fetal bovine serum; MTT, 3-(4,5-dimethylthiazol2-yl)-2,5-diphenyltetrazolium bromide; NGF, nerve growth factor; NVP, 4 4 1-naphthylvinyl)pyridine; rp-HPLC, reverse-phase HPLC; SDS-PAGE, SDS-polyacrylamide gel electrophoresis.
258
TROPHIC FACTOR EFFECTS ON NEUROBLASTOMA CELLS understanding the mechanisms of neuronal differentiation (Bottenstein, 1981). In vitro, these cell lines express neuronal phenotypes and retain the capacity to undergo both biochemical and morphological differentiation in response to chemical agents such as dibutyryl cyclic AMP, retinoic acid, dimethyl sulfoxide, and phorbol esters (Waymire et al., 1978; Ishii et al., 1985;Gotti et al., 1987;Summerhill et al., 1987;Casper and Davies, 1989). Effects of growth factors on the differentiation of neuronal cell lines have also been documented (Ishii et al., 1985; Pavelic and Spaventi, 1987; Reynolds and Perez-Polo, 1989);however, these studies have focused mainly on growth and morphological differentiation, and, aside from demonstrations of effects of nerve growth factor (NGF) and epidermal growth factor (EGF) on the development of tyrosine hydroxylase and ChAT in pheochromocytoma cell lines (Schubert et a]., 1977; Goodman et al., 1980), little is known about factors affecting the expression of neurotransmitter enzymes. Substances that enhance cholinergic development in embryonic spinal cord (Giller et al., 1977; Smith and Appel, 1983; McManaman et al., 19886), and in sympathetic (Swerts et al., 1983) and parasympathetic neurons (Vaca et al., 1985)have been detected in myotube-conditioned medium and extracts of skeletal muscle. However it is not clear if these same factors also increase the cholinergic development in neuronal cell lines. We have recently isolated a factor from rat skeletal muscle, ChAT development factor (CDF), that enhances the ChAT activity of rat spinal cord cultures (McManaman et al., 1988a) and rescues motoneurons from cell death (McManaman et al., 1990).The active factor is an acidic polypeptide with an apparent molecular mass of 22-24 kDa (McManaman et al., 1988~). The physical and biochemical properties of CDF distinguish it from substances in heart and skeletal muscle conditioned media that enhance cholinergicproperties of sympathetic neurons (Weber et al., 1985;Yamamori et al., 1989), as well as from other target-derived factors with effects on cholinergic properties, including NGF, fibroblast growth factors (FGFs), and insulin-like growth factor-1 (IGF- I). The molecular weight of CDF is similar to that of ciliary neurotrophic factor (CNTF) (Lin et al., 1989). However, these two molecules do not appear to be identical, as the PI value for CDF is distinct from that of CNTF (McManaman et al., 1988~; Lin et al., 1989) and skeletal muscle tissue does not appear to express mRNA for CNTF (Stockli et al., 1989). In addition to CDF, we have also shown that skeletal muscle contains basic fibroblast growth factor (bFGF)-like activity, that bFGF enhances the ChAT activity of rat spinal cord cultures, and that the effects of bFGF and CDF on spinal cord ChAT activity are additive (McManaman et al., 1989). In the present studies we investigate the effects of skeletal muscle factors on the development of ChAT activity in cultures of LA-N-2 cells, a cholinergic human neuroblastoma cell line established by Seeger et
259
al. ( 1977).In serum-containing medium, LA-N-2 cells express both veratridine-dependent Na' channels that are inhibited by tetrodotoxin and the ability to synthesize and release acetylcholine (West et al., 1977; Richardson et al., 1989). Furthermore, this cell line does not express detectable levels of tyrosine hydroxylase or the ability to transport y-aminobutyric acid (West et al., 1977). Thus, these cells appear to provide a useful model system to study the effects of exogenous factors on the development of cholinergic properties. Although cyclic nucleotides (Kirshner et al., 1986)and retinoic acid (Casper and Davies, 1989) have been shown to affect the cholinergic expression of neuroblastoma cells, the effect of neurotrophic factors has not previously been studied. We also compare the effects of CDF and bFGF on ChAT levels, in both LAN-2 cells and PC-12 cells, with those of NGF, EGF, and IGF- 1. Our results indicate that neurotrophic factors differentially affect cholinergic cell lines, and that they may act through diverse intracellular mechanisms.
MATERIALS AND METHODS
Materials Leibovitz's L- 15 medium and glutamine were purchased from GIBCO (Grand Island, NY, U.S.A.). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), trypsin, and general chemicals were obtained from Sigma (St. Louis, MO, U.S.A.). Fetal bovine serum was purchased from Hyclone Labs (Logan, UT, U.S.A.) and gentamicin from ESI Pharmaceuticals (Cherry Hill, NJ, U.S.A.). 4-( 1-Naphthylviny1)pyridine (NVP) was obtained from Calbiochem (La Jolla, CA, U.S.A.). [3H]Acetyl-CoAand [3H]thymidinewere obtained from New England Nuclear (Boston, MA, U.S.A.). Plastic culture plates were obtained from Fisher Scientific. NGF was a generous gift from Prof. J. R. Perez-Polo (University of Texas Medical Branch, Galveston, TX, U.S.A.). IGF-1 was obtained from Amgen (Thousand Oaks, CA, U.S.A.), and EGF from Collaborative Research (Bedford, MA, U.S.A.). bFGF was a gift from Dr. F. Fuller, California Biotechnology (Mt. View, CA, U.S.A.). LA-N-2 cells were kindly provided by Dr. R. Seeger (Childrens Hospital, Los Angeles, CA, U.S.A.).
Cell culture LA-N-2 cells were seeded at a density of 200,000 cells/cm2 in plastic (Costar) culture plates and incubated at 37°C in a humidified atmosphere without C02. For studies of the effects of skeletal muscle fractions on ChAT activity, the cells were grown for 3 days without treatment in Leibovitz's L- 15 medium, containing 2 mM glutamine, 0.05 mg/ml gentamicin (LGG), supplemented with 15% fetal bovine serum (LGG15). The medium was then changed to LGG with 5% fetal bovine serum (LGG-5), and the indicated muscle fractions were added. After 3 days, the cells were assayed for ChAT activity. For time course and cell growth studies, LA-N-2 cells were grown continuously in LGG-5 medium and neurotrophic factors were added at the time of plating. The culture medium was changed every other day, at which time the cells were also retreated with trophic factors. At the indicated times the cells were harvested, counted, and assayed for ChAT activity. Cells were passaged in 0.06% trypsin, 0.02% EDTA in Hank's balanced salt solution.
J. Neurochem.. Vol. 57, No. I , 1991
260
J. L. McMANAMAN AND F. G. CRA WFORD
Extraction and purification of CDFs Soluble extracts of skeletal muscle from 2- to 3-week old Sprague-Dawley rats were prepared and fractionated as previously described by McManaman et al. (1988~).In brief, extracts were centrifuged at 100,000 g, the supernatant fraction (100KS) was then acidified with acetic acid to pH 5, and the resulting precipitate (pH 5-P) was removed by centrifugation, resolubilized at pH 9.0, and subjected to gel filtration chromatography on Sephadex G 100. Gel filtration chromatography resulted in both high (GIOOA) and low (G100B) molecular weight activity peaks. The G IOOB fraction was then sequentially fractionated on hydroxylapatite at pH 9.0 (HAP-9), DEAE cellulose chromatography (DEAE), and hydroxylapatite at pH 7.0 (HAP-7). The HAP-7 fraction was then subjected to either preparative sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) (McManaman et al., 19880) or to reverse-phase HPLC (rpHPLC). Purification by rp-HPLC was conducted at room temperature using a Vydac C4 column. The sample (50 pl) was applied to the C4 column and eluted with a linear gradient of 40-80% acetonitrile in 0.1% trifluoroacetic acid over a period of 60 min, at a flow rate of 0.5 ml/min. The purity of CDF was determined by subjecting CDF to SDS-PAGE in 14% acrylamide gels (McManaman et al., 1988a), using the silver staining procedure described by Oakley et al. (1980) to detect protein bands. Protein was assayed by the bicinchoninic acid method (Smith et al., 1985).
ImmunobIot analysis Protein samples were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes (Immobilon-P, Millipore), and reacted with antibodies according to the procedure of Towbin et al. (1979), as previously described by McManaman et al. (1989). A portion of the membrane was removed and stained with Coomassie Blue to compare the protein staining pattern with the immunoreactivity. The immunoreactive bands were visualized by reacting the membranes with peroxidase-conjugated rabbit anti-mouse immunoglobulin G (IgG).
Preparation of antibodies to CDF Monoclonal antibodies to CDF were prepared according to Kohler (1980). BALB/c mice were immunized with 10 pg of SDS-PAGE purified CDF in Freund's complete adjuvant. Spleen cells from immunized animals were fused with SP2 myeloma cells and hybrids were selected by the method of Littlefield (1964). Hybrids producing antibodies to CDF were identified by immunoblot analysis (Towbin et al., 1979) and cloned by limiting dilution.
for 10 min at 37"C, and repeatedly pipetting the cells until they were completely dispersed. The cells were then diluted fourfold in LGG-5, an aliquot was mixed with trypan blue, and both trypan blue-positive and -negative cells were counted using a hemocytometer. Over 90% of the cells harvested by this technique excluded trypan blue. The remainder of the cells were pelleted at 800 g, the excess medium removed by aspiration, and the cells were assayed for ChAT as described above. Control studies documented that the cell number and protein content of LA-N-2 cultures were proportional for all treatments and under all culture conditions. For protein determinations, suspended cells were pelleted at 800 g, and resuspended in 0.01 M Na2HP04,0.15 M NaCl, 0.004 M KCl, pH 7.2. This procedure was repeated twice, and aliquots were then taken for cell counts and protein measurements (Smith et al., 1985) and ChAT assay. In the present studies we have expressed the specific activity of ChAT as femtomoles [3H]acetylcholine synthesized per hour per cell to stress the cellular effects of the trophic factors. However, identical results were obtained expressing ChAT specific activity per milligram of protein.
Metabolic activity The metabolic activity of the cultures was determined, as previously described by McManaman et al. (1990), by the MTT method, which involves the conversion of soluble MTT to an insoluble product by intact cells with actively metabolizing mitochondria (Mosmann, 1983).
Growth factor activities The biological activities of the neurotrophic and growth factors used in these studies were determined from doseresponse curves, using previously established biological assays. The activity of CDF was determined by its ability to enhance ChAT in embryonic day 14 (E14) rat spinal cord cultures (McManaman et al., 1988a). The ED50values of individual lots of CDF averaged 25 k 20 nM (n = 7) and 55 & 30 nM (n = 6) on spinal cord and LA-N-2 cultures, respectively. The activities of IGF-1, bFGF, and EGF were verified by their ability to enhance [3H]thymidine uptake in 3T3 fibroblast cultures (McManaman et al., 1989).The activity of NGF was verified by its ability to enhance neurite outgrowth on PC12 cells (Greene and Rukenstein, 1989). In preliminary experiments we tested a wide range of concentrations ofeach factor on LA-N-2 cultures (data not shown), both to determine their maximal concentrations and to verify their effects.
Statistical analyses Student's t tests were applied to the data for statistical analysis. All statements of significance refer to p < 0.05.
ChAT assay ChAT activity in LA-N-2 cells was measured by a modification of the method of Fonnum ( I 975), which is based on the generation of [3H]acetylcholinefrom [3H]acetyl-CoAand unlabeled choline, as previously described for spinal cord cultures (McManaman et al., 19886). The reaction solution contained 0.1% Nonidet P-40, 5 mM NaH2P04, 150 mM NaC1, 10 mM EDTA, 1.5 m M choline chloride, 0.1 mM eserine sulfate, 0.5 mg/ml bovine serum albumin, and 0.1 mM [3H]acetyl-CoA (sp act, 32 pCi/pmol). NVP was used to verify that the acetyltransferase activity was due to ChAT and not the result of nonspecific acetyltransferase activity (Cavallito et al., 1970). The specific ChAT activity was determined on single cell suspensions of LA-N-2 cells that were prepared by removing the culture medium, treating the cultures with 0.06% trypsin J. Netcrochem., Vol. 57. No I , 1991
RESULTS Extracts of skeletal muscle enhance ChAT development in LA-N-2 cells LA-N-2 cells constitutively express ChAT activity when cultured in Leibovitz's medium containing fetal bovine serum. However, as shown in Fig. 1, the specific ChAT activity (ChAT/cell) of untreated LA-N-2 cultures remains constant throughout the first week of culture. In contrast, the addition of crude extracts of rat skeletal muscle to the cultures markedly increases their specific ChAT activity. Figure 1 shows that the specific ChAT activity of LA-N-2 cells is increased
261
TROPHIC FACTOR EFFECTS ON NEUROBLASTOMA CELLS
TABLE 1. ChAT enhancing activity of skeletal muscle fractions on LA-N-2 cells
900800-
6 700-
Purification step (number of repetitions)
Specific activity (units/mg of protein ? SD)
lOOKS (5) pH 5-P(5) pH 5-S(5) GlOOA (3) GlOOB (3) HAP-9 (2) DEAE (3) HAP-7 13)
30.9 ? 8.1 70.4 f 16.7 84.4 k 84.2 58.4 f 13.6 974.4 ? 365.9 3,026.5 f 271.5 4,022.4 k 232.9 8,057.7 f 2,336.9
2
-.$ 8
600 500-
L
4
400-
3
300-
+
200100~
0'
I
0
1
2
3
4
5
6
I
Culture Time (Days)
FIG. 1. Extracts of rat skeletal muscle enhance the level of ChAT activity in LA-N-2 cultures. LA-N-2 cells were cultured for 6 days in LGG-5 medium in the presence or absence of acid precipitated rat skeletal muscle extracts (pH 5-P). The specific ChAT activities in untreated LA-N-2 cultures (W) were assayed at the indicated times. The effect of 150 fig of pH 5-P fraction on the specific ChAT activity of LA-N-2 cells was assayed after 3 and 6 days of treatment (A).ChAT activity is shown as the percent of the specific activity of cultures assayed at the time of treatment (day 0). Each value is the average of triplicate culture wells (kSD) from a representative time course experiment. This experiment has been repeated four times with similar results.
fivefold on day 3 and ninefold on day 6 by treatment with the soluble fraction (pH 5-P) obtained from acidprecipitated rat skeletal muscle extract (see Materials and Methods). Both the basal and the induced ChAT activities in LA-N-2 cultures are completely (>98%) inhibited by 10 pA4 NVP, a specific inhibitor of ChAT (Cavallito et al., 1970). Thus, the acetyltransferase activity in the cultures is due to ChAT and not to nonspecific transferases. The ChAT development activity on LA-N-2 cells copurifies with CDF Skeletal muscle factors that enhance the development of ChAT activity in embryonic rat spinal cord cells have previously been fractionated by a combination of acid precipitation and chromatography on Sephadex G 100, DEAE cellulose, and hydroxylapatite (McManaman et al., 1988~).As shown in Table 1, this fractionation scheme also results in the enrichment of substances that stimulate ChAT in LA-N-2 cells. In each of these purification steps the activity responsible for stimulating the development of ChAT in LA-N-2 cells had the same elution characteristics as that which increased ChAT in embryonic spinal cord cells (data not shown). We have previously isolated CDF from the HAP-7 fraction by preparative SDS-PAGE (McManaman et al., 1988~).To determine if it is this, or other, factors that are responsible for increasing ChAT in LA-N-2 cells, we compared the electrophoretic migration of factors affecting ChAT development in cultures of E 14 rat spinal cord and LA-N-2 cells. Figure 2 shows that
Skeletal muscle fractions containing substances that enhance ChAT activity in El4 rat spinal cord cells were assayed for the ability to enhance ChAT activity in LA-N-2 cultures. The specific ChAT enhancing activity (units/mg of protein) of each fraction was determined by treating triplicate culture wells with increasing amounts of the specified fraction and assaying for ChAT after 3 days of incubation. Each value is the mean ? SD of multiple purifications of rat skeletal muscle. The identities of the fractions have previously been described in detail by McManaman et al. (1988~)and are briefly defined in Materials and Methods. Activity units are operationally defined as the amount of protein required to increase ChAT activity by 50% over control values (McManaman et al., 1988a).
a single peak of activity is obtained for both types of cultures, and the activity on LA-N-2 cells comigrates with the activity on spinal cord cultures. To provide additional, independent, evidence that it is CDF that is responsible for increasing ChAT in LA-N-2 cells, we purified the ChAT enhancing activity in the HAP-7 fraction by rp-HPLC. Figure 3 shows
=
75k
50k
39k
27k
5
10
15
20
17k
3004
4
G 6
100-
0
0
25
30
35
40
45
-4 50
Slice Number
FIG. 2. Comparison of the preparative SDS-PAGE properties of
ChAT enhancing activity on LA-N-2 and rat spinal cord cultures. ChAT enhancing activity in the HAP-7 fraction was purified by preparativeSDS-PAGE and tested on cultures of LA-N-2 and €14 rat spinal cord neurons. The amount of ChAT enhancing activity in extracts of individual 2-rnm gel slices on LA-N-2 (M)and spinal is shown as a percentage of the activity found cord cultures (0) in untreated cultures. The values are the means of triplicate culture wells treated with the indicated fractions from a representative experiment. This experiment has been repeated three times with identicalresults. The error in the ChAT assay is
Skeletal Muscle Proteins Stimulate Cholinergic Differentiation of Human Neuroblastorna Cells *?$James L. McManaman and *?Frances G. Crawford *Wagner ALS Research Laboratory and t The Jerry Lewis Neuromuscular Disease Research Center, Department of Neurology, Division of Neuroscience, and $Program in Cellular and Molecular Biology, Baylor College of Medicine, Houston, Texas, U S A .
Abstract: Extracts of rat skeletal muscle contain substances that enhance the development of choline acetyltransferase (ChAT) in the cholinergic human neuroblastoma cell line LA-N-2. The ChAT enhancing activity in muscle extract was purified to homogeneity by preparative gel electrophoresis and reverse-phase HPLC. The active factor is biochemically and immunologically identical to ChAT development factor, (CDF), the skeletal muscle factor that enhances ChAT activity in enriched cultures of embryonic rat motoneurons and rescues motoneurons from naturally occumng cell death in vivo. CDF increases the specific ChAT activity of LA-N-2 cells fivefold after 6 days in culture, but does not affect their growth or metabolic activity. Basic fibroblast growth factor also in-
creases ChAT activity in LA-N-2 cells and its effect is additive with that of CDF. In contrast, neither insulin-like growth factor- I , epidermal growth factor, nor nerve growth factor affected the ChAT activity of LA-N-2 cells. Our study demonstrates for the first time that CDF can directly affect the development of neuronal properties in a homogeneous population of cells, and that the effects of CDF are separate from those of other types of trophic factors. Key Words: Muscle proteins-Purification-Neuronal development-cholinergic enzymes-Neurotrophic factors-Neuroblastoma cells. McManaman J. L. and Crawford F. G. Skeletal muscle proteins stimulate cholinergic differentiation of human neuroblastoma cells. J . Neurochem. 57, 258-266 (199 1).
The functional differentiation of cholinergic neurons is associated with an increase in the levels of choline acetyltransferase (ChAT), the enzyme responsible for the synthesis of acetylcholine (Burt, 1975; Pilar et al., 1981). Although the factors that regulate the expression of this enzyme are, for the most part, unknown, increasing evidence supports the involvement of exogenous target-derived factors. This evidence is based primarily on the effects of isolated neurotrophic factors and tissue extracts on the development of ChAT activity in primary cultures of embryonic neurons (Hefti
et al., 1985; McManaman et al., 1988a,b; Saadat et al., 1989; Yamamori et al., 1989). Interpretation of these studies, however, has been complicated both by the heterogeneous nature of primary cultures and by effects of target-derived factors on neuronal survival (Martinou et al., 1989; McManaman et al., 1990). Use of cholinergic cell lines provides a means of investigating the effects of neurotrophic factors on the development of cholinergic properties in a homogeneous population of cells, in which survival is not a complicating factor. Neuronal cell lines are important model systems for
Received October 9, 1990; revised manuscript received December 20. 1990; accepted December 2 1, 1990. Address correspondence and reprint requests to Dr. J. McManaman at Department of Neurology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, U.S.A. Abbreviations used: bFGF, basic fibroblast growth factor; CDF, ChAT development factor; ChAT, choline acetyltransferase; DEAE, biologically active fraction obtained following chromatography on DEAE cellulose; E, embryonic day; EGF, epidermal growth factor; GI OOA, Sephadex G 100 void volume fraction that contains a high molecular weight form of ChAT enhancing activity;GIOOB, Sephadex
G 100 fraction that contains CDF; HAP-7, biologically active fraction obtained following chromatography on hydroxylapatite at pH 7.0; HAP-9, biologically active fraction obtained following chromatography on hydroxylapatite at pH 9.0; IGF-1, insulin-like growth factorI ; IgG, immunoglobulin G; LGG-5, identical to LGG-15, but containing 5% fetal bovine serum; LGG-15, Leibovitz's L-15 medium containing 2 mM glutamine, 0.05 mg/ml gentamicin, and supplemented with 15% fetal bovine serum; MTT, 3-(4,5-dimethylthiazol2-yl)-2,5-diphenyltetrazolium bromide; NGF, nerve growth factor; NVP, 4 4 1-naphthylvinyl)pyridine; rp-HPLC, reverse-phase HPLC; SDS-PAGE, SDS-polyacrylamide gel electrophoresis.
258
TROPHIC FACTOR EFFECTS ON NEUROBLASTOMA CELLS understanding the mechanisms of neuronal differentiation (Bottenstein, 1981). In vitro, these cell lines express neuronal phenotypes and retain the capacity to undergo both biochemical and morphological differentiation in response to chemical agents such as dibutyryl cyclic AMP, retinoic acid, dimethyl sulfoxide, and phorbol esters (Waymire et al., 1978; Ishii et al., 1985;Gotti et al., 1987;Summerhill et al., 1987;Casper and Davies, 1989). Effects of growth factors on the differentiation of neuronal cell lines have also been documented (Ishii et al., 1985; Pavelic and Spaventi, 1987; Reynolds and Perez-Polo, 1989);however, these studies have focused mainly on growth and morphological differentiation, and, aside from demonstrations of effects of nerve growth factor (NGF) and epidermal growth factor (EGF) on the development of tyrosine hydroxylase and ChAT in pheochromocytoma cell lines (Schubert et a]., 1977; Goodman et al., 1980), little is known about factors affecting the expression of neurotransmitter enzymes. Substances that enhance cholinergic development in embryonic spinal cord (Giller et al., 1977; Smith and Appel, 1983; McManaman et al., 19886), and in sympathetic (Swerts et al., 1983) and parasympathetic neurons (Vaca et al., 1985)have been detected in myotube-conditioned medium and extracts of skeletal muscle. However it is not clear if these same factors also increase the cholinergic development in neuronal cell lines. We have recently isolated a factor from rat skeletal muscle, ChAT development factor (CDF), that enhances the ChAT activity of rat spinal cord cultures (McManaman et al., 1988a) and rescues motoneurons from cell death (McManaman et al., 1990).The active factor is an acidic polypeptide with an apparent molecular mass of 22-24 kDa (McManaman et al., 1988~). The physical and biochemical properties of CDF distinguish it from substances in heart and skeletal muscle conditioned media that enhance cholinergicproperties of sympathetic neurons (Weber et al., 1985;Yamamori et al., 1989), as well as from other target-derived factors with effects on cholinergic properties, including NGF, fibroblast growth factors (FGFs), and insulin-like growth factor-1 (IGF- I). The molecular weight of CDF is similar to that of ciliary neurotrophic factor (CNTF) (Lin et al., 1989). However, these two molecules do not appear to be identical, as the PI value for CDF is distinct from that of CNTF (McManaman et al., 1988~; Lin et al., 1989) and skeletal muscle tissue does not appear to express mRNA for CNTF (Stockli et al., 1989). In addition to CDF, we have also shown that skeletal muscle contains basic fibroblast growth factor (bFGF)-like activity, that bFGF enhances the ChAT activity of rat spinal cord cultures, and that the effects of bFGF and CDF on spinal cord ChAT activity are additive (McManaman et al., 1989). In the present studies we investigate the effects of skeletal muscle factors on the development of ChAT activity in cultures of LA-N-2 cells, a cholinergic human neuroblastoma cell line established by Seeger et
259
al. ( 1977).In serum-containing medium, LA-N-2 cells express both veratridine-dependent Na' channels that are inhibited by tetrodotoxin and the ability to synthesize and release acetylcholine (West et al., 1977; Richardson et al., 1989). Furthermore, this cell line does not express detectable levels of tyrosine hydroxylase or the ability to transport y-aminobutyric acid (West et al., 1977). Thus, these cells appear to provide a useful model system to study the effects of exogenous factors on the development of cholinergic properties. Although cyclic nucleotides (Kirshner et al., 1986)and retinoic acid (Casper and Davies, 1989) have been shown to affect the cholinergic expression of neuroblastoma cells, the effect of neurotrophic factors has not previously been studied. We also compare the effects of CDF and bFGF on ChAT levels, in both LAN-2 cells and PC-12 cells, with those of NGF, EGF, and IGF- 1. Our results indicate that neurotrophic factors differentially affect cholinergic cell lines, and that they may act through diverse intracellular mechanisms.
MATERIALS AND METHODS
Materials Leibovitz's L- 15 medium and glutamine were purchased from GIBCO (Grand Island, NY, U.S.A.). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), trypsin, and general chemicals were obtained from Sigma (St. Louis, MO, U.S.A.). Fetal bovine serum was purchased from Hyclone Labs (Logan, UT, U.S.A.) and gentamicin from ESI Pharmaceuticals (Cherry Hill, NJ, U.S.A.). 4-( 1-Naphthylviny1)pyridine (NVP) was obtained from Calbiochem (La Jolla, CA, U.S.A.). [3H]Acetyl-CoAand [3H]thymidinewere obtained from New England Nuclear (Boston, MA, U.S.A.). Plastic culture plates were obtained from Fisher Scientific. NGF was a generous gift from Prof. J. R. Perez-Polo (University of Texas Medical Branch, Galveston, TX, U.S.A.). IGF-1 was obtained from Amgen (Thousand Oaks, CA, U.S.A.), and EGF from Collaborative Research (Bedford, MA, U.S.A.). bFGF was a gift from Dr. F. Fuller, California Biotechnology (Mt. View, CA, U.S.A.). LA-N-2 cells were kindly provided by Dr. R. Seeger (Childrens Hospital, Los Angeles, CA, U.S.A.).
Cell culture LA-N-2 cells were seeded at a density of 200,000 cells/cm2 in plastic (Costar) culture plates and incubated at 37°C in a humidified atmosphere without C02. For studies of the effects of skeletal muscle fractions on ChAT activity, the cells were grown for 3 days without treatment in Leibovitz's L- 15 medium, containing 2 mM glutamine, 0.05 mg/ml gentamicin (LGG), supplemented with 15% fetal bovine serum (LGG15). The medium was then changed to LGG with 5% fetal bovine serum (LGG-5), and the indicated muscle fractions were added. After 3 days, the cells were assayed for ChAT activity. For time course and cell growth studies, LA-N-2 cells were grown continuously in LGG-5 medium and neurotrophic factors were added at the time of plating. The culture medium was changed every other day, at which time the cells were also retreated with trophic factors. At the indicated times the cells were harvested, counted, and assayed for ChAT activity. Cells were passaged in 0.06% trypsin, 0.02% EDTA in Hank's balanced salt solution.
J. Neurochem.. Vol. 57, No. I , 1991
260
J. L. McMANAMAN AND F. G. CRA WFORD
Extraction and purification of CDFs Soluble extracts of skeletal muscle from 2- to 3-week old Sprague-Dawley rats were prepared and fractionated as previously described by McManaman et al. (1988~).In brief, extracts were centrifuged at 100,000 g, the supernatant fraction (100KS) was then acidified with acetic acid to pH 5, and the resulting precipitate (pH 5-P) was removed by centrifugation, resolubilized at pH 9.0, and subjected to gel filtration chromatography on Sephadex G 100. Gel filtration chromatography resulted in both high (GIOOA) and low (G100B) molecular weight activity peaks. The G IOOB fraction was then sequentially fractionated on hydroxylapatite at pH 9.0 (HAP-9), DEAE cellulose chromatography (DEAE), and hydroxylapatite at pH 7.0 (HAP-7). The HAP-7 fraction was then subjected to either preparative sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) (McManaman et al., 19880) or to reverse-phase HPLC (rpHPLC). Purification by rp-HPLC was conducted at room temperature using a Vydac C4 column. The sample (50 pl) was applied to the C4 column and eluted with a linear gradient of 40-80% acetonitrile in 0.1% trifluoroacetic acid over a period of 60 min, at a flow rate of 0.5 ml/min. The purity of CDF was determined by subjecting CDF to SDS-PAGE in 14% acrylamide gels (McManaman et al., 1988a), using the silver staining procedure described by Oakley et al. (1980) to detect protein bands. Protein was assayed by the bicinchoninic acid method (Smith et al., 1985).
ImmunobIot analysis Protein samples were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes (Immobilon-P, Millipore), and reacted with antibodies according to the procedure of Towbin et al. (1979), as previously described by McManaman et al. (1989). A portion of the membrane was removed and stained with Coomassie Blue to compare the protein staining pattern with the immunoreactivity. The immunoreactive bands were visualized by reacting the membranes with peroxidase-conjugated rabbit anti-mouse immunoglobulin G (IgG).
Preparation of antibodies to CDF Monoclonal antibodies to CDF were prepared according to Kohler (1980). BALB/c mice were immunized with 10 pg of SDS-PAGE purified CDF in Freund's complete adjuvant. Spleen cells from immunized animals were fused with SP2 myeloma cells and hybrids were selected by the method of Littlefield (1964). Hybrids producing antibodies to CDF were identified by immunoblot analysis (Towbin et al., 1979) and cloned by limiting dilution.
for 10 min at 37"C, and repeatedly pipetting the cells until they were completely dispersed. The cells were then diluted fourfold in LGG-5, an aliquot was mixed with trypan blue, and both trypan blue-positive and -negative cells were counted using a hemocytometer. Over 90% of the cells harvested by this technique excluded trypan blue. The remainder of the cells were pelleted at 800 g, the excess medium removed by aspiration, and the cells were assayed for ChAT as described above. Control studies documented that the cell number and protein content of LA-N-2 cultures were proportional for all treatments and under all culture conditions. For protein determinations, suspended cells were pelleted at 800 g, and resuspended in 0.01 M Na2HP04,0.15 M NaCl, 0.004 M KCl, pH 7.2. This procedure was repeated twice, and aliquots were then taken for cell counts and protein measurements (Smith et al., 1985) and ChAT assay. In the present studies we have expressed the specific activity of ChAT as femtomoles [3H]acetylcholine synthesized per hour per cell to stress the cellular effects of the trophic factors. However, identical results were obtained expressing ChAT specific activity per milligram of protein.
Metabolic activity The metabolic activity of the cultures was determined, as previously described by McManaman et al. (1990), by the MTT method, which involves the conversion of soluble MTT to an insoluble product by intact cells with actively metabolizing mitochondria (Mosmann, 1983).
Growth factor activities The biological activities of the neurotrophic and growth factors used in these studies were determined from doseresponse curves, using previously established biological assays. The activity of CDF was determined by its ability to enhance ChAT in embryonic day 14 (E14) rat spinal cord cultures (McManaman et al., 1988a). The ED50values of individual lots of CDF averaged 25 k 20 nM (n = 7) and 55 & 30 nM (n = 6) on spinal cord and LA-N-2 cultures, respectively. The activities of IGF-1, bFGF, and EGF were verified by their ability to enhance [3H]thymidine uptake in 3T3 fibroblast cultures (McManaman et al., 1989).The activity of NGF was verified by its ability to enhance neurite outgrowth on PC12 cells (Greene and Rukenstein, 1989). In preliminary experiments we tested a wide range of concentrations ofeach factor on LA-N-2 cultures (data not shown), both to determine their maximal concentrations and to verify their effects.
Statistical analyses Student's t tests were applied to the data for statistical analysis. All statements of significance refer to p < 0.05.
ChAT assay ChAT activity in LA-N-2 cells was measured by a modification of the method of Fonnum ( I 975), which is based on the generation of [3H]acetylcholinefrom [3H]acetyl-CoAand unlabeled choline, as previously described for spinal cord cultures (McManaman et al., 19886). The reaction solution contained 0.1% Nonidet P-40, 5 mM NaH2P04, 150 mM NaC1, 10 mM EDTA, 1.5 m M choline chloride, 0.1 mM eserine sulfate, 0.5 mg/ml bovine serum albumin, and 0.1 mM [3H]acetyl-CoA (sp act, 32 pCi/pmol). NVP was used to verify that the acetyltransferase activity was due to ChAT and not the result of nonspecific acetyltransferase activity (Cavallito et al., 1970). The specific ChAT activity was determined on single cell suspensions of LA-N-2 cells that were prepared by removing the culture medium, treating the cultures with 0.06% trypsin J. Netcrochem., Vol. 57. No I , 1991
RESULTS Extracts of skeletal muscle enhance ChAT development in LA-N-2 cells LA-N-2 cells constitutively express ChAT activity when cultured in Leibovitz's medium containing fetal bovine serum. However, as shown in Fig. 1, the specific ChAT activity (ChAT/cell) of untreated LA-N-2 cultures remains constant throughout the first week of culture. In contrast, the addition of crude extracts of rat skeletal muscle to the cultures markedly increases their specific ChAT activity. Figure 1 shows that the specific ChAT activity of LA-N-2 cells is increased
261
TROPHIC FACTOR EFFECTS ON NEUROBLASTOMA CELLS
TABLE 1. ChAT enhancing activity of skeletal muscle fractions on LA-N-2 cells
900800-
6 700-
Purification step (number of repetitions)
Specific activity (units/mg of protein ? SD)
lOOKS (5) pH 5-P(5) pH 5-S(5) GlOOA (3) GlOOB (3) HAP-9 (2) DEAE (3) HAP-7 13)
30.9 ? 8.1 70.4 f 16.7 84.4 k 84.2 58.4 f 13.6 974.4 ? 365.9 3,026.5 f 271.5 4,022.4 k 232.9 8,057.7 f 2,336.9
2
-.$ 8
600 500-
L
4
400-
3
300-
+
200100~
0'
I
0
1
2
3
4
5
6
I
Culture Time (Days)
FIG. 1. Extracts of rat skeletal muscle enhance the level of ChAT activity in LA-N-2 cultures. LA-N-2 cells were cultured for 6 days in LGG-5 medium in the presence or absence of acid precipitated rat skeletal muscle extracts (pH 5-P). The specific ChAT activities in untreated LA-N-2 cultures (W) were assayed at the indicated times. The effect of 150 fig of pH 5-P fraction on the specific ChAT activity of LA-N-2 cells was assayed after 3 and 6 days of treatment (A).ChAT activity is shown as the percent of the specific activity of cultures assayed at the time of treatment (day 0). Each value is the average of triplicate culture wells (kSD) from a representative time course experiment. This experiment has been repeated four times with similar results.
fivefold on day 3 and ninefold on day 6 by treatment with the soluble fraction (pH 5-P) obtained from acidprecipitated rat skeletal muscle extract (see Materials and Methods). Both the basal and the induced ChAT activities in LA-N-2 cultures are completely (>98%) inhibited by 10 pA4 NVP, a specific inhibitor of ChAT (Cavallito et al., 1970). Thus, the acetyltransferase activity in the cultures is due to ChAT and not to nonspecific transferases. The ChAT development activity on LA-N-2 cells copurifies with CDF Skeletal muscle factors that enhance the development of ChAT activity in embryonic rat spinal cord cells have previously been fractionated by a combination of acid precipitation and chromatography on Sephadex G 100, DEAE cellulose, and hydroxylapatite (McManaman et al., 1988~).As shown in Table 1, this fractionation scheme also results in the enrichment of substances that stimulate ChAT in LA-N-2 cells. In each of these purification steps the activity responsible for stimulating the development of ChAT in LA-N-2 cells had the same elution characteristics as that which increased ChAT in embryonic spinal cord cells (data not shown). We have previously isolated CDF from the HAP-7 fraction by preparative SDS-PAGE (McManaman et al., 1988~).To determine if it is this, or other, factors that are responsible for increasing ChAT in LA-N-2 cells, we compared the electrophoretic migration of factors affecting ChAT development in cultures of E 14 rat spinal cord and LA-N-2 cells. Figure 2 shows that
Skeletal muscle fractions containing substances that enhance ChAT activity in El4 rat spinal cord cells were assayed for the ability to enhance ChAT activity in LA-N-2 cultures. The specific ChAT enhancing activity (units/mg of protein) of each fraction was determined by treating triplicate culture wells with increasing amounts of the specified fraction and assaying for ChAT after 3 days of incubation. Each value is the mean ? SD of multiple purifications of rat skeletal muscle. The identities of the fractions have previously been described in detail by McManaman et al. (1988~)and are briefly defined in Materials and Methods. Activity units are operationally defined as the amount of protein required to increase ChAT activity by 50% over control values (McManaman et al., 1988a).
a single peak of activity is obtained for both types of cultures, and the activity on LA-N-2 cells comigrates with the activity on spinal cord cultures. To provide additional, independent, evidence that it is CDF that is responsible for increasing ChAT in LA-N-2 cells, we purified the ChAT enhancing activity in the HAP-7 fraction by rp-HPLC. Figure 3 shows
=
75k
50k
39k
27k
5
10
15
20
17k
3004
4
G 6
100-
0
0
25
30
35
40
45
-4 50
Slice Number
FIG. 2. Comparison of the preparative SDS-PAGE properties of
ChAT enhancing activity on LA-N-2 and rat spinal cord cultures. ChAT enhancing activity in the HAP-7 fraction was purified by preparativeSDS-PAGE and tested on cultures of LA-N-2 and €14 rat spinal cord neurons. The amount of ChAT enhancing activity in extracts of individual 2-rnm gel slices on LA-N-2 (M)and spinal is shown as a percentage of the activity found cord cultures (0) in untreated cultures. The values are the means of triplicate culture wells treated with the indicated fractions from a representative experiment. This experiment has been repeated three times with identicalresults. The error in the ChAT assay is
