sign in sign up
<<
Home , Insulin, Nerve growth factor

The Journal of Neuroscience May 1986, 6(5): 1211-1219

Effects of , Insulin-like -II, and on Formation and Survival in Cultured Sympathetic and Sensory

Esperanza Recio-Pinto,’ Matthew M. Rechler,* and Douglas N. Ishii’f *Laboratory of Biochemical Pharmacology, National Institute of Arthritis, , and Digestive Diseases, National Institutes of Health, Bethesda, Maryland 20205, and TDepartment of Physiology and Biophysics, Colorado State University, Fort Collins, Colorado 80523

Insulin and the insulin-like growth factors (IGFs) may directly pathetic (Collins and Dawson, 1983) neurons. With respect to affect the development of the . NGF, IGF-II, and the CNS, ornithine decarboxylase activity (Roger and Fellows, insulin’s effects on neurite formation and neuronal survival were 1980) and uridine and thymidine uptake (Raizada et al., 1980) studied in peripheral ganglion cell cultures from chick embryos. are increased. uptake is reported as increased (Phillip Neurite outgrowth was enhanced in a dose-dependent manner and Coxon, 1976) or unchanged (Goodner and Berrie, 1977). by insulin and IGF-II in sympathetic cell cultures. The half- Furthermore, insulin binding sites have been detected in brain maximally effective concentration, ED,,, was about 0.4-0.6 nM homogenates (Havrankova et al., 1978a; Pacold and Blackard, for both polypeptides, and concentrations as low as 10 PM were 1979) and plasma membrane fractions (Landau et al., 1983; active. However, in sensory neurons the ED,, for neurite out- Posner et al., 1974). growth was about 30 nM for insulin and 0.1 nM for IGF-II, The observations cited above have not stimulated quite the suggesting that these factors may have selective effects in dif- same intense investigative effort as in the NGF field, perhaps ferent neuronal tissues. Neither serum nor the presence of non- because of the following reservations: (1) Supraphysiological neuronal cells was required for the response in sympathetic neu- insulin concentrations were studied, raising questions as to phys- rons. The specific anti-NGF antiserum inhibited the neurite out- iological relevance and the possibility that the effects were spu- growth response to NGF but not to insulin nor IGF-II. Insulin rious, e.g., due to an active contaminant. (2) The increased and IGF-II additionally supported survival of sensory and sym- incorporation of precurser molecules could be predominantly pathetic neurons; however, insulin was not as efficacious as NGF. into the non-neuronal cells. (3) Insulin may not be acting directly The combination of high concentrations of NGF and insulin was on the neurons. For example, the activity of ganglion satellite no better than NGF alone in supporting sympathetic cell sur- cells can substantially influence neuronal survival (Bumham et vival, or neurite outgrowth. This indicates that insulin acts on al., 1972). NGF, in addition, can be produced by various non- the same, or a subpopulation, of NGF-responsive neurons. These neuronal cells (Murphy et al., 1977; Young et al., 1975). There- results support the hypothesis that insulin and its homologs fore, insulin may only indirectly enhance neuronal survival and belong to a broad family of neuritogenic polypeptides. neurite formation through stimulation of satellite cells. (4) The function of brain binding sites is unsettled, partly because sites Other than NGF, few, if any, chemically well-defined and phys- are present on both neurons and glia (Raizada et al., 1982). Glia iologically important factors are known to be capable of en- comprise the large majority of brain cells. At issue also is the hancing neurite outgrowth (Barde et al., 1983; Mobley et al., problem of how insulin can reach beyond the few privileged 1977). In the presence of , the role of the sites in brain, in the face of the blood-brain barrier. Conse- substrate in supporting neurite outgrowth has been shown for quently, a clear-cut demonstration that physiological concen- substances such as (Baron-Van Evercooren et al., 1982; trations of insulin can act directly on neurons could help over- Manthorpe et al., 1983). The identification of new neuritogenic come many of these reservations. agents would advance the present understanding of the devel- Recently, we found that concentrations as low as 1O-l 00 PM opment of the nervous system. We suggest that insulin and its of insulin and insulin-like growth factor-II (IGF-II) could en- homologs may belong to a broad family of neuritogenic poly- hance neurite outgrowth in the cloned neuroblastoma . Although it has been the traditional view that neurons SH-SYSY cell line (Recio-Pinto and Ishii, 1984; Recio-Pinto et are not target cells for insulin, the mounting evidence suggests al., 1984). Non-neurons were clearly unnecessary for the re- that insulin acts on both the PNS and CNS. For example, in sponse. The implication that insulin and its homologs have an the PNS, insulin treatment of cultured sensory ganglia can in- important direct role in the nervous system is further supported crease uridine and incorporation (Burnham et al., 1974; by the observation that the coding for IGF-II is expressed Levi-Montalcini, 1964). Neurite outgrowth is enhanced in sen- in brain (Soares et al., 1985). IGF-II is closely related in structure sory (Bothwell, 1982; Snyder and Kim, 1980) and parasym- to insulin (Rinderknecht and Humbel, 1978) and has overlap- ping activities (Rinderknecht and Humbel, 1976; Zapf et al., Received Oct. 11, 1984; revised Nov. 1, 1985; accepted Nov. 1, 1985. 1978). Moreover, (D’Ercole et al., 1980) and in- We thank Gregory M. Ishii and Aaron T. Ishii for fine technical assistance in sulinlike (Havrankova et al., 1978b) activities have been de- parts of this study. This work was supported by Grant RO 1 AM3284 1 from the tected in brain. In this study, we have considered whether (1) National Institute of Arthritis, , Digestive and Diseases. insulin and IGF-II can enhance neurite formation and survival Correspondence should be addressed to Dr. Ishii, Physiology and Biophysics Department, Colorado State University, Fort Collins, CO 80523. in sensory and sympathetic cultures; (2) physiological concen- 1 Present address: Department of Physiology and Anesthesiology, Cornell Uni- trations are active; and (3) the effects are directly on the neuronal versity Medical Center, 1300 York Ave., New York, NY 10021. cells. We report that physiological concentrations of insulin and Copyright 0 1986 Society for Neuroscience 0270-6474/86/05 121 l-09$02.00/0 IGF-II are directly neuritogenic and share certain other key

1211 1212 Recio-Pinto et al. Vol. 6, No. 5, May 1986

horn chick embryos were dissociated and cells cultured in the plating medium described by Mains and Patterson (1973), which was modified in the following way. The L- 15 plating medium contained their vitamin mix, 3 mg/ml BSA, and 0.1 mg/ml transfenin,but NGF wasomitted. Fetal calf serum (5 or 10%) was present only where indicated. cell cultures Dorsal root ganglia from IO-d-old chick embryos were dissociated, and the cells were cultured in the plating medium comprised of the RPM1

( 1640 medium supplemented with the N2 mixture described by Bottenstein and Sato (1979), but without insulin. Other conditions 0 16" l$ 10' 13 0 1P 1iP 1iP aredescribed in the legends. insulin [MI NGF IMI Cytosine arabinofuranosidetreatment of cultures Figure 1. Effect of various concentrations of insulin and NGF on neu- In some experiments with sympathetic cells, cytosine arabinofuranoside rite outgrowth in sympathetic ganglion cells cultured in serum-contain- (ara-C) was used to inhibit the growth of non-neuronal cells (Wood and ing medium. The cells were cultured for 1 d in the L- 15 plating medium Bunge, 1975). The cells were incubated for 2 d in the plating medium wtih 10% fetal calf serum and the indicated concentrations of insulin which contained, in addition, 12% fetal calf serum and 20 rig/ml NGF or NGF. Neurite outgrowth was scored. Two separate experiments are or insulin as described. Cytosine arabinofuranoside (2.5 &ml) was shown: triangles,12-d-old chick embryos; circles, 13-d-old embryos. added and the incubations continued for 3 more d, at which time very The filled circle shows the effect of NGF together with 1 PM insulin. few non-neuronal cells were visible. The cultures were washed and the The values are means f SEM; n = 4 replicate cultures. cells were incubated further as described. properties with NGF, a neurotrophic agent whose influence on Neurite outgrowth and neuronal survival the developmentof the sensoryand sympathetic nervoussystem The characteristic morphology of chick sympathetic (Charnley et al., is well documented(Levi-Montalcini and Angeletti, 1968; Mob- 1972; Mains and Patterson, 1973) and sensory (Scott et al., 1969) gan- ley et al., 1977). glion cell cultures have been described. For example, the capsule, neu- rilemma, and fibroblast cells are 7-10 pm in diameter and assume a Materials and Methods fibroblast-like shape after 1 d in culture (Scott et al., 1969). But the sensory neurons are phase-bright, about 30 Nrn in diameter, and remain Materials round in shape. Sympathetic neurons are more variable in size. For our Pork pancreatic insulin (24 U/mg) was purchased from Sigma Chemical purposes, the spherical cells that were bright and had a clear nucleus Co. (St. Louis). It was dissolved in 0.01 N HCl and stored at -20°C. under modulation contrast were scored as neurons. The scoring of neu- IGF-II is also known as rat cell multiplication-stimulating activity rites followed a procedure similar to that used previously (Sonnenfeld (Acquaviva et al., 1982). The particular IGF-II preparation used is and Ishii, 1982; Spinelli et al., 1982). Briefly, the percentage of neurons designated MSA-111-2, which is 7 100 in molecular weight and has been bearing was determined by replicate counts on more than 100 purified to homogeneity (Moses et al., 1980). The beta subunit of NGF neurons in several randomly chosen fields under low-power modulation was prepared from male mouse (Burton et al., 1978). Its purity, contrast microscopy, A field was 1.2% of the total culture area. Where confirmed by the presence of a single band on gel electrophoresis, and survival was poor, additional fields were counted so that at least 100 biological potency are described elsewhere (Ishii, 1978). The anti-NGF neurons were always scored. Only neurons with processes equal to or antis&m was prepared by injecting the toe pads of 6- to I-week-old longer than 40 pm were scored as positive. Neurons with multiple pro- female New Zealand albino rabbits with 2 ma of the uurified beta NGF cesses were scored only once. Neurites could generally be identified by in the presence of complete Freund’s adjuvant. The antiserum gave a the presence of growth cones. To score survival, the neurons found in single precipitin band against purified NGF and against submaxillary four 100X fieldsin predeterminedlocations along the north-southand 105,000 x g supernatant on Ouchterlony plates and immuno- east-west axes were counted in each of four replicate cultures. Values electrophoresis. The preparation and properties of the antiserum are are means and SEM. more completely described elsewhere (Ishii and Shooter, 1975). The antiserum was twice precipitated at 4°C with 45% saturated ammonium Statistical analysis sulfate that had been neutralized with ammonium hydroxide. The re- Analysis of variance and the Bonferroni multiple t test (Dunn and Clark, sulting crude gamma globulin fraction was dialyzed against 0.15 M NaCl 1974) were used to compare experimental groups. Data expressed as and 10 mM KH,PO,, pH 7, and stored frozen in aliquots at -20°C. percentageswere transformed to arcsinevalues prior to statisticalanal- Methocel (65HG, 4000 cp, Fuka AG) was purchased from Tridon ysis. Chemical Co. (Hauppauge, NY). Ganglion cell cultures Results The excised ganglia were washed in Roswell Park Memorial Institute Efects of insulin and NGF on neurite outgrowth in Culture Medium 1640 (RPM1 1640) and incubated for 20 min at 37°C sympathetic cell cultures in Hanks’ salt solution containing 3 1 U/ml trypsin for sympathetic and 6.2 U/ml for dorsal root ganglia. The solution was discarded. In serum-containingmedium, the maximum proportion of neu- The ganglia were washed 3 x in RPM1 1640 medium with 10% serum, rite-bearing cells is observed within a day, although the neurites resuspended in medium of the same composition, and dissociated by do continue to elongate. The sympathetic ganglion cells were trituration through a Pasteur pipette. Debris was removed by permitting cultured in the presenceof various concentrations of insulin or undissociated fragments to settle for 2 min, and the cells were collected NGF for 1 d; then the proportion of neuronswith neuriteswas by centrifugation at 600 x g for 5 min. The cell pellet was resuspended scored. The neurite outgrowth responseto insulin from two in the appropriate plating medium described below, 50-85,000 cells separate experiments is shown in Figure 1A. The ED,,, the were seeded in 1.O ml onto sulfate-coated 17.6 mm multiwell plates, and incubated at 37°C in humidified air under various conditions. concentration at which the half-maximal neurite outgrowth re- The cultures were fed on alternate days. The tissue culture plates were sponseis obtained, was about 0.4 nM for each experiment. The prepared in advance by treatment for at least 2 hr in 1 mg/ml protamine contribution of insulin from the serum, about 4 PM, was neg- sulfate, followed by washing 3 x with RPM1 1640 medium. ligible and therefore not included in the concentrations shown in Figure 1A. Note that the same sensitivity is observed in Sympathetic ganglion cell cultures serum-free medium (Fig. 2). The data obtained from cultures Several methods were evaluated, and the following was the most sat- with and without serum were pooled, and the ED,, was about isfactory. Lumbar sympathetic ganglia from 12- to 13-d-old white leg- 0.57 f 0.19 nM (mean f SEM, n = 3 experiments). An exten- The Journal of Neuroscience Insulin, IGF-II, and NGF Effects on Neurite Formation 1213 sive network of fasciculated neurites, which covered the entire culture surface, was observed after several weeks. In the sameexperiment indicated by the circles in Figure lA, the responseto NGF was studied (Fig. 1B). The ED,, was about 0.1 &ml (4 PM), a value closeto that previously reported (Ishii, T T 1 1978). The maximum responsewas about 77%, similar to the value obtained by Varon and Raiborn (1972). Two observations suggestthat insulin may be acting on the same or a subpopu- lation of NGF-responsive neurons. First, the sameproportion of sympathetic neuronswas responsiveto insulin (Fig. IA) and to NGF (Fig. 1B). Second, the combination of high concentra- tions of insulin and NGF had no greatereffect than either factor alone (Fig. 1B). The experiments on survival describedbelow provide additional strong support for this hypothesis. The culture medium containedtransferrin, which can increase neuronal survival (Bottenstein, 1980). In the absenceof trans- ferrin, the sensitivity to insulin, basedon the ED,,, fell to about 20 nM from 0.4 nM. The maximum neurite outgrowth response varied somewhatfrom experiment to experiment.

Effect of IGF-II on neurite outgrowth in Figure 2. Effect of variousconcentrations of IGF-11 and insulinon sympathetic cell cultures neuriteoutgrowth in sympatheticganglion cells cultured in serum-free The effect of various concentrations of IGF-II and insulin on medium.The cellswere cultured for 3.5 d in the L- 15 platingmedium neurite outgrowth in cells cultured in serum-free medium is without serumand with the indicatedconcentrations of IGF-11(0) or shown in Figure 2. The slower rate and apparent lower maxi- insulin (0). Neurite outgrowthwas scored.The valuesare meansf mum in the neurite outgrowth responseto insulin under the SEM; n = 3 replicatecultures. serum-freecondition compared to that in serum-containingme- dium is due, perhaps, to a slower metabolic rate under the former condition. Insulin and IGF-II concentrations as low as morphology of IGF-II-treated cultures could not be distin- 10 PM were effective. The ED,, for neurite outgrowth was about guished from that of insulin-treated cultures (not shown). 0.4 nM for both polypeptides. The IGF-II dose-responsecurve was about the samewhen additionally studied in serum-con- Effect of insulin and IGF-II on neurite outgrowth taining cultures (not shown). Evidently, the other components in sensorycell cultures in serum are not required for the neurite outgrowth responseto Dorsal root ganglion cells were cultured in serum-freemedium insulin and IGF-II. The level of spontaneousneurite outgrowth with various concentrations of insulin and IGF-II; neurite out- was lower in serum-free (Fig. 2) than in serum-containingme- growth was scoredafter 3.5 d (Fig. 4). IGF-II concentrationsas dium (Fig. l), which indicatesthat the endogenousinsulin, IGF- low as a few picomoles were active. The ED,, for neurite out- II, NGF, and other neuritogenic factors ordinarily present in growth was about 30 nM for insulin and 0.1 nM for IGF-II. serum are active. Spontaneousneurite outgrowth is sensitiveto These resultsshow that physiological concentrations of IGF-II serum factors in SH-SYSY cells also (Recio-Pinto et al., 1984). can enhanceneurite outgrowth in sensoryas well as in sym- pathetic neurons. Moreover, there appearsto be a tissue-de- Efect of anti-NGF antiserum on the neurite outgrowth pendent difference in sensitivity to these factors. Insulin and responseto insulin, IGF-II, and NGF in sympathetic IGF-II are equally potent in sympathetic neurons (Fig. 2), but cell cultures insulin is about 300-fold less potent than IGF-II in sensory Studies were conducted to determine whether NGF might be present as a contaminant in preparations of insulin and IGF- II. The effect of the specific anti-NGF antiserum (Ishii and Table 1. Effect of the anti-NGF antiserum on neurite outgrowth Shooter, 1975)on the neurite outgrowth responsewas examined. mediated by insulin, IGF-11, and NGF in embryonic chick Insulin, IGF-II, and NGF increasedneurite outgrowth (Table sympathetic neurons 1). The anti-NGF antiserum inhibited the neurite outgrowth responseto NGF but not to insulin or IGF-II. The amount of Neurite outgrowth anti-NGF antiserum usedwas sufficient to neutralize the insulin Additions (%) preparation, even were all of it NGF. The other studies that None 17.8 5 1.8 showthat the effectsof insulin and IGF-II are distinct from that Anti-NGF, 10 pi/ml of NGF are discussedbelow. 29.8 k 4.80 NGF, 1 rig/ml 82.7 + 5.7h Effect of insulin, IGF-II, and NGF on the morphology NGF, 1 &ml plusanti-NGF, 10 &ml 17.6 + 70 of cultured sympathetic cells Insulin, 10 nM 65.1 f 5.6h Sympathetic cells were cultured in serum-containing medium Insulin, 10 nM plusanti-NGF, 10&ml 67.0 f 7.3Q,h supplementedwith insulin, NGF, or both factors for 2 d. The IGF-11, 1 nM 80.1 f 6.4b morphology of the culturesis shown in Figure 3. In the absence IGF-11, 1 nM plus anti-NGF, 10 pi/ml 82.4 f 4%+ of treatment, very few neurites were observed,and the number of neurons was greatly diminished (Fig. 3A). However, with Sympatheticganglion cells from 12-d-old embryonic chicks were cultured for 1 d in medium containing serum and the indicated additions. The cultures were insulin and NGF treatment, many neurites and more neurons scored for the proportion of net&e-bearing cells. Values are means + SEM; n = were seen(Fig. 3, B, c). After a few weeks,all cultures showed 4 replicate cultures. a dense, interconnecting network of neuronal processes.A re- ENot significantly different from group immediately preceding. markable difference in the morphology between insulin- and bp < 0.01 between group and untreated cultures. NGF-treated cultures was not noticed. In other studies, the c p < 0.01 between group and group immediately preceding. 1214 Recio-Pinto et al. Vol. 6, No. 5, May 1966

Figure 3. Morphologyof sympatheticcells cultured in the L- 15plating medium with 10%serum, insulin, and NGF for 2 d: A, untreated;B, plus 1 FM insulin;C, plus 0.4 nM NGF; and D, plus 1 PM insulin and0.4 nM NGF. Abbreviations:N, ;S, non-neuron.Modulation contrast, x 260. neurons(Fig. 4). This appearsto be due to a lower sensitivity Efect of insulin, IGF-II, and NGF on the survival of to insulin in the sensoryneurons from 1O-d-old embryos. Like embryonic neurons the sympatheticneurons, sensory neurons did not require serum NGF is an important requirement for the survival of sensory for the responseto insulin or IGF-II. and sympathetic neurons(for review, seeMobley et al., 1977). The Journal of Neuroscience Insulin, IGF-II, and NGF Effects on Neurite Formation 1215

80 c iz 5 3 60 - 0 & 3 O 40 - Q, c ; Insulin IMI NGF IMI z 20 - Figure 5. Effect of various concentrations of insulin and NGF on sur- vival in cultured sympathetic neurons. The experiment is a continuation of that described in Figure 1 and the conditions and the symbols are the same, except that at 2.5 d the number of surviving neurons in four -12 -IO -8 predetermined fields in each culture was counted (82,000 cells were 0 10 10 10 12 plated). A field is 1.2% of the total culture area. Values are means f Polypeptide (M) SEM; n = 4 replicate cultures. Figure 4. Effect of various concentrations of IGF-II and insulin on neurite outgrowth in dorsal root ganglion cells. Sensory cell cultures Effect of insulin on neurite outgrowth in sympathetic from 1O-d-old embryonic chicks were cultured in the RPM1 1640 plating culturesfollowing treatment with Methocel or medium without serum and with various concentrations of IGF-II (0) cytosine arabinoside and insulin (0). Neurite outgrowth was scored after 3.5 d. The values It was possiblethat insulin was acting indirectly through effects are means+ range;n = 2 replicatecultures (100-150 cells scored per on non-neurons.Therefore, sympathetic ganglioncells were cul- culture). tured under two separateconditions known to inhibit the pro- liferation of non-neurons. Mains and Patterson (1973) have shown that sympathetic neurons may be cultured free of non- neuronalcells by deletion of bicarbonate and addition of Meth- The effect of various concentrations of insulin and NGF on the ocel. Our culture medium was a slight modification. Protamine number of surviving sympathetic neuronsis shown in Figure 5. sulfate was used instead of as the substrate,and 0.1 We canestimate the number of surviving neurons.The maximal mg/ml transfenin and 10% fetal calf serum were included. Fetal survival responseto NGF, basedon Figure 5B, is approximately calf serum is toxic to rat (Mains and Patterson, 1973) but not (700 neurons/area of four fields) x (total area of culture) = to chick sympathetic neurons. The morphology of chick sym- 14,600 neurons. The number of total cells seededwas 82,000. pathetic ganglion cells cultured for 2 d in this medium is shown About half of the seededcells at this stageof development are in Figure 6. The non-neurons ordinarily present (Fig. 3) are neurons(Varon and Raibom, 1972). Therefore, we calculate a virtually absent. The survival of neurons was poor, and few plating efficiency of 36% for the neurons. This value may be neurites were observed in control cultures (Fig. 6A). Survival comparedwith the 25% observed by Varon and Raibom (1972) and neurite outgrowth were enhancedby insulin, NGF, and their under their culture conditions. NGF appeared to be more effi- combination under conditions in which the non-neuronal cells caciousthan insulin in supporting neuronal survival. If NGF were virtually absent (Fig. 6, B-D). and insulin were acting on separateneuronal cell populations, Cytosine arabinofuranoside (ara-C) is toxic to proliferating treatment with both factors should result in an additive effect non-neuronal ganglion cells but not to the essentiallyquiescent on survival. However, the combination of high concentrations neuronal cells. After scoring of the experiment shown by the of insulin and NGF had no greater effect than NGF alone (Fig. triangles in Figure lA, the cultures were treated with 2.5 pg/ml SB). The interesting implications will be discussed. ara-C for 4 d. Thereafter, the cultures were washed, and the These resultswere obtained as part of the sameexperiment incubations were continued in the same concentrations of in- shown in Figure 1 (indicated by the circles), and the two effects sulin as before. Even after 3 weeks, non-neuronal cells were may be compared. The dose-responsecurves for the effects of absent, whereasa continuous monolayer of mesenchymecells insulin on neurite outgrowth (Fig. 1A) and survival (Fig. 54) was ordinarily observed in cultures not treated with ara-C. appearto sharethe samesensitivity. The sameseems to be the Nevertheless,the neuritesbecame very long in the insulin-treat- casefor NGF (Figs. lB, 5B). Survival and neurite outgrowth, ed cultures, forming networks like those following long-term then, are functions that share the same dose-responsecurve. treatment with NGF. However, these functions need not be linked. Consider that These results show that non-neuronal cells are neither re- neurite outgrowth was scored at 1 d (Fig. 1A) and survival at quired for the effectsof insulin on neurite outgrowth nor survival 2.5 d (Fig. 5). Neurons, of course, do not immediately die in in sympathetic neurons. the absenceof thesefactors. It is, then, not surprisingthat 100% of surviving neurons do not have neurites, particularly early Discussion after seeding.In fact, Varon and Raibom (1972) have observed Our results show that physiological concentrations of insulin separatetime courses in the survival and neurite outgrowth and IGF-II can enhancesurvival and neurite formation in sym- responsesto NGF in sympathetic neurons.Survival and neurite pathetic and sensoryneurons. These effectsare observed with outgrowth are clearly distinct responsesalso in neuroblastoma insulin even in sympathetic cell cultures that are essentially cells (Sonnenfeld and Ishii, 1982). devoid of non-neuronal cells. It is evident then that insulin and Attempts were not made to quantify the survival effect of IGF-II do act directly on neurons. We shall now discussthe IGF-II, but it was visually apparent that IGF-II was a survival implications of the effectson neuronal survival, the neuritogenic factor too. Furthermore, insulin and IGF-II enhancedthe sur- role of thesefactors in different cell types, the validity of these vival of sensoryneurons. findings, and the mechanismof action. 1216 Recio-Pinto et al. Vol. 6, No. 5, May 1986

Figure 6. Effectof Methocelon nenriteoutgrowth and the morphologyof sympatheticcells cultured in mediumwith insulinand NGF. Cellswere culturedfor 2 d in the L- 15plating medium containing 7 mg/ml Methocel,10% serum, and the indicatedadditions: A, no additions;B, plus2 pM insulin;C, plus0.4 nM NGF, D, plus 2 PM insulinand 0.4 nM NGF. Figurelabel: N, neuron.Modulation contrast, x 260.The nenriteoutgrowth in a separateexperiment, under these conditions in whichthe non-nenronalcells were virtually eliminated,was as follows: no additions,11.2 -C 1.5%;plus insulin, 42.8 +-6.0%; plus NGF, 58.0f 6.0%;plus insulin and NGF, 66.8 rt 7.2. The valuesare means k SEW,12 = 4 replicatecultures. Survival plays an important role in sculpting the nervous preprogrammeddeath of overproduced neuronal cells. During systemduring development (reviewed in Mobley et al., 1977). synaptogenesisit appearsthat there is competition among grow- For example, the hyperplasiaof sympatheticganglia that follows ing neurites for the target cell sourceof NGF, which undergoes treatment with NGF is largely the result of a decreasein the retrograde axonal transport and modulates neuronal cell func- The Journal of Neuroscience Insulin, IGF-II, and NGF Effects on Neurite Formation 1217

Table 2. Half-maximally effective concentrations (ED,,) for neurite outgrowth stimulated by insulin and IGF-II in different types of cells

Neurite outgrowth (ED,,) Insulin IGF-II Type of cell b4 ml Reference Chick parasympathetic ciliary neurons 160 Collins and Dawson, 1983 Chick sensory neurons 50 4 Bothwell, 1982 30 0.1 This study Chick sympathetic neurons 0.4 0.4 This study Human neuroblastoma cells 4 0.5 Recio-Pinto and Ishii, 1984 O.la Recio-Pinto and Ishii, 1984 u Studied in the presence of the protease inhibitor bacitracin. tion. It is probable that other factors, in dddition to NGF, are stimulate neurite formation in different types of cells are shown important to survival. In the absenceof a high concentration in Table 2. Insulin and IGF-II are equally potent in sympathetic of transferrin, NGF alone is insufficient to support survival of neurons (Fig. 2). However, insulin is 300-fold lesspotent than sympathetic and sensory neurons and serum is generally in- IGF-II in sensory neurons (Fig. 4). The selective responsein- cluded in cultures. It is, therefore, noteworthy that low concen- dicates that insulin may play a role subordinate to that of the trations of insulin and IGF-II can enhance survival (Fig. 5). IGFs in lo-d-old embryonic sensoryneurons. Our ED,, value This is a direct effect. Insulin supports survival of about half as for insulin is similar to that reported by Bothwell (1982). The many neuronsas NGF, and appearsless efficacious. The com- small difference may be due to the presenceof transfen-inin our bination of optimum concentrations of both factors results in cultures. The larger difference in ED,, values for IGF-II is prob- no greater survival than with NGF alone. This suggestseither ably the result, at least in part, of Bothwell’s useof commercial that insulin acts on a subpopulationof the NGF-responsive cells IGF-II, which we too have found to be variable in potency from or that insulin is less efficacious and can support survival of lot to lot. High concentrations of insulin are also required to only a fraction of all of the NGF-responsive neurons. It is also stimulate neurite formation in parasympathetic neurons(Col- possiblethat there is no difference in the efficacy of insulin and lins and Dawson, 1983). As with the sensoryneurons, insulin NGF in supporting the survival of attached neurons.Cultures may not be the preferred . initiated under NGF-treatment do not appear to suffer a sub- The validity of thesefindings requiressome comment. These stantial loss of neuronsfollowing subsequentwashout of NGF resultsare not attributable to NGF as a contaminant in insulin and continued treatment with insulin alone. This suggeststhat and IGF-II preparations. The anti-NGF antiserum does not insulin may not be as active as NGF in supporting neurons inhibit the responseto insulin or IGF-II (Table 1). On the other during their critical transition into culture, possibly on events hand, the anti-insulin antiserum doesinhibit neurite outgrowth such as attachment. Whatever the mechanism,the poorer sur- stimulated by insulin, but not NGF, in SH-SYSY cells (Recio- vival with insulin posesa problem for the experimentalist. The Pinto et al., 1984). Preparations of insulin and IGF-II do not improved survival of neuronsin transferrin-supplementedcul- compete for the binding of 1251-NGFto its specifichigh-affinity tures probably explains, in large measure, why we have been receptors(Recio-Pinto and Ishii, 1984). Purified human insulin, able to observe these effects at physiological rather than the prepared by bacteria1 synthesis of A and B chains, and com- supraphysiologicalconcentrations studied by others. bination of the chains,is aspotent asporcine and bovine Neurite outgrowth is potentially a secondaryconsequence of in enhancing neurite outgrowth in human neuroblastomacells. the effectsof insulin on survival (Bothwell, 1982; Snyder and It is highly unlikely that synthetic human insulin and insulin Kim, 1980). This view appearsto be supported by the obser- extracted from tissue sourceshave equal amounts of the same vation that the dose-responsecurves for neurite outgrowth (Fig. active contaminant. Moreover, there is a qualitative difference 1A) and survival (Fig. 5A) are very similar. The neurons are in the activity of insulin and NGF preparations. Unlike NGF, exposedto NGF and other factors prior to explantation. Survival insulin by itself cannot enhanceneurite outgrowth in cultured enhancedby insulin may simply permit a transient expression PC12 cells (Recio-Pinto et al., 1984). in culture of the stimulation obtained in vivo. We believe this It is difficult to study the mechanismof IGF-II and insulin- to be an implausible explanation for insulin’s activity. Insulin dependent neurite formation in sensoryand sympathetic neu- and IGF-II can support neurites in cells cultured for as long as rons, becausethese factors alsosupport neuronalsurvival. Dying 3 weeks,and it is very unlikely that neuritogenic factors active untreated cultures do not provide adequateexperimental con- prior to cultivation have as long a residual effect. Survival and trols. The human neuroblastomaSH-SYSY cell is an alternative neurite outgrowth need not be linked, asalready shownfor NGF model for the study of neuritogenesis(Ishii et al., 1985), useful in neurons (Campenot, 1977; Gundersen and Barrett, 1980; partly becauseit doesnot require thesefactors for survival (Son- Menesini Chen et al., 1978). They are also unlinked in the nenfeld and Ishii, 1982). Most likely, neuroblastomais derived responseof human neuroblastomacells to NGF (Sonnenfeld from sympathetic cells, and possibly from sensorycells. SH- and Ishii, 1982), insulin (Recio-Pinto and Ishii, 1984), and tu- SYSY cells contain tyrosine hydroxylase and beta- mor promoters (Spinelli et al., 1982). However, additional stud- hydroxylase (Biedler et al., 1978), indicating their sympathetic ies are necessaryto demonstrate directly that these responses phenotype. Insulin and IGF-II directly and reversibly enhance are unlinked when induced by insulin and IGF-II in sympathetic neurite formation in SH-SYSY cells (Recio-Pinto and Ishii, and dorsal root ganglia cells. 1984). The ED,, values in cells cultured with and without the There can be substantialcross-occupancy of insulin and IGF proteaseinhibitor bacitracin are shown in Table 2. Insulin and receptors(Rechler et al., 1983). Insulin’s low potency suggests IGF-II have similar potencies in human SH-SYSY and chick that it acts through the IGF sitesin both sensoryand parasym- sympathetic cells. Serum is required in neither of these kinds pathetic neurons.The half-maximal effective concentrationsthat of cells for the neurite outgrowth response.Microtubules are Recio-Pinto et al. Vol. 6, No. 5, May 1986 important cytoskeletal elementsof . There is a good cor- tissues to take up labeled insulin in vivo and to respond to insulin in relation between insulin’s capacity to increasetubulin mRNA vitro. 101: 605-6 12. levels and enhanceneurite outgrowth in SH-SYSY cells (Mill Gundersen, R. W., and J. N. Barrett (1980) Characterization of the et al., 1985). IGF-II also stimulates an increasein tran- turning response of dorsal root neurites towards nerve growth factor. J. Cell Biol. 87: 546-554. script levels. Insulin increases3H-leucine and 3H-uridine uptake Havrankova, J., J. Roth, and M. Brownstein (1978a) Insulin receptors (Recio-Pinto and Ishii, 1984). These increases,and neurite out- are widely distributed in the central nervous system of the rat. Nature growth, are inhibited by cycloheximide and actinomycin D, 272: 827-829. respectively. The effect of cycloheximide is reversible. These Havrankova, J., D. Schmechel, J. Roth, and M. Brownstein (1978b) observations suggestthat synthesisis important in in- Identification of insulin in rat brain. Proc. Natl. Acad. Sci. USA 75: sulin-stimulatedneurite formation. NGF (Sonnenfeldand Ishii, 5737-5741. 1982) and tumor promoters (Spinelli et al., 1982) are other Ishii, D. N. (1978) Effect of tumor promoters on the response of significant factors that stimulate neurite formation in SH-SYSY cultured embryonic chick ganglia to nerve growth factor. Cancer Res. cells. It is unlikely that insulin is acting directly through the 38: 3886-3893. Ishii, D. N., and E. M. Shooter (1975) Regulation of nerve growth NGF or tumor- receptors, for insulin doesnot inhibit factor synthesis in mouse submaxillary by . J. Neu- tumor-promoter binding (Spinelli and Ishii, 1983), and neither rochem. 25: 843-85 1. insulin nor IGF-II inhibits NGF binding (Recio-Pinto and Ishii, Ishii, D. N., E. Recio-Pinto, W. Spinelli, J. E. Mill, and K. H. Sonnenfeld 1984). Separateinsulin and IGF receptorshave been identified (1985) Neuriteformation modulated by nervegrowth factor, insulin, on thesecells (unpublished observations). and tumor promoter receptors. Int. J. Neurosci. 26: 109-127. The contents of this report, when taken together with the Landau, B. R., Y. Takaoka, M. A. Abrams, S. M. Genuth, M. van observations of others, support the reasonablehypothesis that Houten, B. I. Posner, R. J. White, S. Ohgaku, A. Horvat, and E. insulin and IGF-II are membersof a broad family of structurally Hemmelgam (1983) Binding of insulin by monkey and hypo- thalamus. Diabetes 32: 284-292. related neuritogenic agents.These factors are but two of a series Levi-Montalcini, R. (1964) The nerve growth factor: Its mode ofaction of structurally related polypeptides with overlapping actions on sensory and sympatheticnerve cells.Harvey Lect. 60: 2 17-259. (Rinderknecht and Humbel, 1976; Zapf et al., 1978). It seems Levi-Montalcini, R., and P. U. Angeletti (1968) Nerve growth factor. likely that other memberswill be uncovered. Physiol. Rev. 48: 534-569. Mains, R. E., and P. H. Patterson (1973) Primary cultures of disso- ciated sympathetic neurons. I. Establishment of long-term growth in References culture and studies of differentiated properties. J. Cell Biol. 59: 329- Acquaviva, A. M., C. B. Bruni, S. P. Nissley, and M. M. Rechler (1982) 345. Cell-free synthesis of rat insulin-like growth factor II. Diabetes 31: Manthorpe, M., E. Engvall, E. Ruoslahti, F. M. Longo, G. E. Davis, 656-658. and S. Varon (1983) Laminin promotes neuritic regeneration from Barde, Y.-A., D. Edgar, and H. Thoenen (1983) New neurotrophic cultured peripheral and central neurons. J. Cell Biol. 97: 1882-l 890. factors. Annu. Rev. Physiol. 45: 601-612. Menesini Chen, M. G., J. S. Chen, and R. Levi-Montalcini (1978) Baron-Van Evercooren, A., H. K. Kleinman, S. Ohno, P. Marangos, J. Sympathetic nerve fibers ingrowth in the central nervous system of P. Schwartz, and M. E. Dubois-Dalcq (1982) Nerve growth factor, neonatal rodent upon intracerebral NGF . Arch. Ital. Biol. laminin, and promote neurite growth in human fetal sen- 116: 53-84. sory ganglia cultures. J. Neurosci. Res. 8: 179-l 93. Mill, J. F., M. V. Chao, and D. N. Ishii (1985) Insulin, insulinlike Biedler, J. L., S. R. Tarlov, M. Schachner, and L S. Freedman (1978) growth factor-II, and nerve growth factor effects on tubulin mRNA Multiple neurotransmitter synthesis by human neuroblastoma cell levels and neuriteformation. Proc. Natl. Acad. Sci. USA 82: 7 126- lines and clones. Cancer Res. 38: 375 l-3757. 7130. Bothwell, M. (1982) Insulin and somatomedin MSA promote nerve Mobley, W. C., A. C. Server, D. N. Ishii, R. J. Riopelle, and E. M. growth factor-independent neurite formation by cultured chick dorsal Shooter (1977) Nerve growth factor. N. Engl. J. Med. 297: 1096- root ganglionic sensory neurons. J. Neurosci. Res. 8: 225-23 1. 1104 (Part I), 1149-l 158 (Part II), and 1211-1218 (Part III). Bottenstein, J. E. (1980) Serum-free culture of neuroblastoma cells. Moses, A. C., S. P. Nissley, P. A. Short, M. M. Rechler, and J. M. In Advances in Neuroblastoma Research, Audrey E. Evans, ed., pp. Podskalny (1980) Purification and characterization of multiplica- 16 l-l 70, Raven, New York. tion-stimulating activity. Insulin-like growth factors purified from rat Bottenstein, J. E., and G. H. Sato (1979) Growth of a rat neuroblas- toma cell line in serum-free supplemented medium. Proc. Natl. Acad. liver cell conditioned medium. Eur. J. Biochem. 103: 387-400. Sci. USA 76: 514-517. Murphy, R. A., R. H. Singer, J. D. Saide, N. J. Pantazi, M. H. Blanchard, Bumham. P. A.. C. Raibom. and S. Varon (1972) Replacement of K. S. Byron, B. G. W. Amason, and M. Young. (1977) Synthesis nerve growth factor by ganglionic non-neuronal cells for the survival and of a high molecular weight form of nerve growth factor in vitro of dissociated ganglionic neurons. Proc. Natl. Acad. Sci. USA by cells in culture. Proc. Natl. Acad. Sci. USA 74: 69: 3556-3560. 4496-4500. Bumham, P. A., J. A. Silva, and S. Varon (1974) Anabolic responses Pacold, S. T., and W. G. Blackard (1979) Central nervous system of embryonic dorsal root ganglia to nerve growth factor, insulin, con- insulin receptors in normal and diabetic rats. Endocrinology 105: canavalin A or serum in vitro. J. Neurochem. 23: 689-695. 1452-1457. Burton. L. E., W. H. Wilson, and E. M. Shooter (1978) Nerve growth Phillip, M. E., and R. V. Coxon (1976) Effect of insulin and pheno- factor in mouse saliva: Rapid isolation procedures for and charac- barbital on uptake of 2-deoxyglucose by brain slices and hemidia- terisation of 7s nerve arowth factor. J. Biol. Chem. 253: 7807-78 12. phragms. J. Neurochem. 27: 643-645. Camuenot. R. B. (1977)-Local control ofneurite development by nerve Posner, B. I., P. A. Kelly, R. P. C. Shiu, and H. G. Friesen (1974) growth factor. Proc. Natl. Acad. Sci. USA 74: 45 16-4519. Studies of insulin, and binding: Tissue Chamlev. J. H.. G. E. Mark. G. R. Campbell, and G. Bumstock (1972) distribution, variation and characterization. Endocrinoloev_a Sympathetic’ganglia in culture. Z. Z&for&h. 135: 287-3 14. 95: 521-531. Collins, F., and A. Dawson (1983) An effect of nerve growth factor Raizada, M. K., J. W. Yang, and R. E. Fellows (1980) Binding of lzsI- on parasympathetic neurite outgrowth. Proc. Natl. Acad. Sci. USA insulin to specific receptors and stimulation of nucleotide incorpo- 80: 209 l-2094. ration in cells cultured from rat brain, Brain Res. 200: 389-400. D’Ercole, A. J., G. T. Applewhite, and L. E. Underwood (1980) Evi- Raizada, M. K., J. F. Stamler, J. T. Quinlan, S. Landas, and M. I. dence that somatomedin is synthesized by multiple tissues in the fetus. Phillips (1982) Identification of insulin -containing cells in Dev. Biol. 75: 315-328. primary cultures of rat brain. Cell. Mol. Neurobiol. 2: 47-52. Dunn, 0. J., and V. A. Clark (1974) Applied Statistics: Analysis of Rechler, M. M., M. Kasuga, N. Sasaki, M. A. de Vroede, J. A. Romanus, Variance and Regression, Wiley, New York. and S. P. Nissley (1983) Properties of insulin-like growth factor Goodner, C. J., and M. A. Berrie (1977) The failure ofrat hypothalamic receptor subtypes. In Insulin-like Growth FactorslSomatomedin Basic The Journal of Neuroscience Insulin, IGF-II, and NGF Effects on Neurite Formation 1219

Chemistry, , and Clinical Importance, E. M. Spencer, ed., pp. and tissue-specific expression of a family of transcripts related to rat 459-490, de Gruyter, New York. insulin-like growth factor II mRNA. Nucleic Acids Res. 13: 1119- Recio-Pinto, E., and D. N. Ishii (1984) Effects of insulin, insulin-like 1134. growth factor-II and nerve growth factor on neurite outgrowth in Sonnenfeld, K. H., and D. N. Ishii (1982) Nerve growth factor effects cultured human neuroblastoma cells. Brain Res. 302: 323-334. and receptors in cultured human neuroblastoma cell lines. J. Neurosci. Recio-Pinto, E., F. F. Lang, and D. N. Ishii (1984) Insulin and in- Res. 8: 375-39 1. sulinlike growth factor-II permit nerve growth factor binding and the Spinelli, W., and D. N. Ishii (1983) Tumor promoter receptors reg- neurite formation response in cultured human neuroblastoma cells. ulating neurite formation in cultured human neuroblastoma cells. Proc. Natl. Acad. Sci. USA 81: 2562-2566. Cancer Res. 43: 4119-4125. Rinderknecht, E., and R. E. Humbel (1976) Polypeptides with non- Spinelli, W., K. H. Sonnenfeld, and D. N. Ishii (1982) Effects of suppressible insulin-like and promoting activities in hu- phorbol tumor promotors and nerve growth factor on neurite man serum: Isolation, chemical characterization, and some biological outgrowth in cultured human neuroblastoma cells. Cancer Res. 42: properties of forms I and II. Proc. Natl. Acad. Sci. USA 73: 2365- 5067-5073. 2369. Varon, S., and P. Raibom (1972) Dissociation, fractionation, and Rinderknecht, E., and R. E. Humbel (1978) Primary structure of hu- culture of chick embryo sympathetic ganglionic cells. J. Neurocytol. man insulin-like growth factor-II. Fed. Eur. Biochem. Sot. Lett. 89: I: 211-221. 283-286. Wood, P. M., and R. P. Bunge (1975) Evidence that sensory axons Roger, L. J., and R. E. Fellows (1980) Stimulation of omithine de- are mitogenic for Schwann cells. Nature 256: 662-664. carboxylase activity by insulin in developing rat brain. Endocrinology Young, M., J. Oger, M. H. Blanchard, H. Asdourian, H. Amos, and B. 106: 619-625. G. W. Amason (1975) Secretion of nerve growth factor by primary Scott, B. S., V. E. Engelberg, and K. C. Fisher (1969) Morphological chick tibroblast cultures. Science 187: 361-362. and electrophysiological characteristics of dissociated chick embry- Zapf, J., E. Schoenle, and E. R. Froesch (1978) Insulin-like growth onic spinal ganglion cells in culture. Exp. Neurol. 23: 230-248. factors I and II. Some biological actions and receptor binding char- Snyder, E. Y., and S. U. Kim (1980) Insulin: Is it a nerve survival acteristics of two purified constituents of nonsuppressible insulin-like factor? Brain Res. 196: 565-571. activity of human serum. Eur. J. Biochem. 87: 285-296. Soares, M. C., D. N. Ishii, and A. Efstratiadis (1985) Developmental