Three distinct neuronal phenotypes exist in embryonic rat hippocampal neurons cultured in basic fibroblast growth factor

ELSEVIER

Neuroseience Letters 204 (1996)5-8

Ntunosu[IgE LHI[RS

Three distinct neuronal phenotypes exist in embryonic rat hippocampal neurons cultured in basic fibroblast growth factor James H. Eubanks a, Jose L. Perez-Velazquez a,b Robert G. Kerr~,a Peter L. Carlen a,b,d,e,Linda R. Mills a,d, Owen T. J o n e s a,b,c,* aplayfair Neuroscience Unit, Toronto Western Hospital, 399 Bathurst Street, Toronto, Ontario M5T 2S8, Canada bBloorview Epilepsy Program, University of Toronto, Toronto, Ontario, Canada CDepartment of Pharmacology, University of Toronto, Toronto, Ontario, Canada dDepartment of Physiology, University of Toronto, Toronto, Ontario, Canada eDepartment of Medicine, University of Toronto, Toronto, Ontario, Canada

Received 21 September1995; revised version received 14 December1995; accepted 14 December1995

Abstract

The possibility that neurons cultured in basic fibroblast growth factor (bFGF) are heterogeneous raises concerns about their subsequent use in gene transfection and transplantation studies. We have examined the fate of embryonic hippocampal neurons in bFGF culture, and now conclude that these cells are not only heterogeneous, but possess neurons of various stages of development. Morphological and immunocytochemical analysis reveal three distinct developmental classes of neurons are present in extended bFGF culture. This tripartite classification is supported by electrophysiological analysis, which reveals that upon depolarization, neurons with precursor and juvenile neuron morphologies are unable to fire action potentials. The third class of neurons, which resemble age-matched polarized neurons in standard serum culture, fired multiple action potentials indicative of a mature phenotype. These data show neurons at multiple developmental st:ages co-exist in bFGF culture, and provide an experimental basis for their classification. Keywords: Growth factors; Development; Hippocampal neurons; Primary culture; Immunocytochemistry; Electrophysiology

Growth factors, acting alone or in concert, orchestrate neuronal development [1,6,11] and are thus promising tools for gene therapeutics and neuronal transplantation [6,9,12,14,15,17]. Especially significant are reports that basic fibroblast growth factor (bFGF, FGF-2) maintains neuronal progenitors in a proliferative mode [6,14,15,18], and that bFGF-treated cells regrow into mature, nontumorigenic cells when subsequently transplanted into brain [6,15]. However, many issues must be addressed before such cells can be used safely in transplantation. A major concern is that the developmental and cellular heterogeneity of the embryonic tissues used for the progenitor cultures may afford transplants with ill-defined, irreproducible or aberrant growth characteristics [6]. Indeed, we recently found that E l 8 hippocampal cultures maintained in bFGF at mitogenic concentrations contain other * Corresponding author. "lbl: +1 416 3695039; fax: +1 416 3695745; e-mail: [email protected].

cell types besides proliferating neuronal precursors [5]. Heterogeneity is especially evident when bFGF cultures are transferred into serum, as cells which arise include maturing neurons, glia, and cells with an ameboid morphology which express neuronal markers [5]. By combining immunocytochemistry and electrophysiology, we now show that bFGF exerts heterologous effects on developing rat hippocampal cultures, by maintaining three distinct developmental stages of neuronal subtypes. Hippocampal cultures were grown on polyornitine/ laminin coated coverslips in N2-bFGF (>4 days) as described [5,14,15]. Cells were fixed with 4% paraformaldehyde in PBS, permeabilized with 0.25% Triton X-100 in PBS, and incubated with antineurofilament antibodies [16]. Following overnight incubation, the cells were washed with PBS and incubated with FITC conjugated goat anti-mouse secondary antibody. Cells were counterstained with 1 ng/ml propidium iodide to define total cells (neurons plus glia), and evaluated by fluorescence mi-

0304-3940/96/$12.00 © 1996 Elsevier Science Ireland Ltd. All fights reserved PII: S0304-3940(96)12337-3

6

J.H. Eubanks et al./Neuroscience Letters 204 (1996) 5-8

croscopy. Many cells, including those which were neurofilament-positive, were found to be engaged in mitosis at the time o f fixation (Fig. 1), but only NF-positive neurons with progenitor morphologies. In no instance were polarized neurons observed in the process of mitosing. Antibodies against microtubule associated protein-2 (MAP-2), a somato-dendritic marker [3], proved to be especially informative indices of neuronal heterogeneity in bFGF. While neurons in standard cultures (age-matched cultures in serum-containing medium) exhibited strong uniform M A P - 2 staining, neurons in b F G F had discreet immunoreactivities which corresponded to each of three morphologically distinct phenotypes we designated as classes 1 3. Class 1 neurons, representing the majority of cells (85% at day 7 and >90% at day 25), were small (at day 7, soma width ( S W 7 ) = 8 . 6 3 ± 0 . 6 6 / a m ; soma length (SL7) = 10.52 ± 0.97/am (n = 14)), non-polarized, weakly M A P - 2 immunoreactive (average pixel intensity

Fig. 2. MAP-2 staining of neurons in bFGF and standard serum cultures. (A) The morphology of neurons at day 7 in standard serum culture. (B-D) The morphologies of MAP-2 reactive class 1, 2, and 3 neurons at day 7 in serum-free N2 bFGF culture. Cultures were counterstained with 1 ng/ml propidium iodide to reveal the nuclei of all cells, both MAP-2 positive and negative that were present in the field. The intensities of staining illustrated in the panels are not normalized, and therefore not intended to provide any quantitative implications between standard and bFGF cultured neurons. The staining intensity values for this comparison are described in the text for each class.

Fig. 1. Mitosis of a neuronal and a non-neuronal cell at day 9 in culture of El8 hippocampal neurons maintained in 10 ng/ml bFGF. Cells were fixed with 4% paraformaldehyde, permeabilized with 0.25% Triton X100 in PBS, and hybridized with anti-neurofilament antibodies. Specific antibody binding was detected using FITC conjugated goat antimouse IgG, and visualized by fluorescence microscopy. The nuclei of all cells were counterstained with propidium iodide. Arrows denote two independent cells caught in the process of mitosis, as may be seen by the propidium iodide staining of the chromatin material in both cells. Neurofilament staining is observed surrounding the propidium iodide stained chromosomes on only one of the two mitosing cells (NF+). No neurofilament staining is observed on the non-neuronal cell caught in the process of cell division (NF-), or on any other cell body in the field. Examples of propidium iodide nuclei staining of the additional cells in the field are illustrated by asterisks on representatives.

at day 7 ( A P 7 ) = 110-120), with very limited processes (Fig. 2B). Class 2 neurons (10% at both days 7 and 25), were somewhat larger (SW7 = 9.0 ± 1.2/am; SL7 = 11.83 ± 0.76/~m (n = 15)), stained more intensely (APT = 126-151) than those o f class 1, and had a few small branched neurites (Fig. 2C). Class 3 cells were the largest neurons (SW7 = 10.0 ± 0.5/am; SL7 = 17.5 _ 1.8/am (n = 4)), possessing extensive neuritic arbors and strong M A P - 2 immunoreactivity (AP7 = 251-300) (Fig. 2D), but were the most infrequent in prevalence (5% at day 7 and <1% at day 25). These class 3 neurons were indistinguishable from mature neurons in standard culture (e.g. SW7 = 11.49 ± 1.05/am; SL7 = 15.83 ± 1.73/am (n = 7)) (Fig. 2A,D). W e next examined the electrophysiological characteristics of the three classes of b F G F neurons disclosed by MAP-2 staining using whole-cell patch clamp techniques [5,9,13]. Standard cultures were again used as a reference phenotype, as cells die in N2 medium lacking b F G F [5,7]. Owing to their preponderance and small size, class 1 neurons were readily identified using our differential contrast interference (DIC) optics. Class 1 neurons rarely fired action potentials even on extreme depolarization (Fig. 3B,

J.H. Eubanks et a l . / Neuroscience Letters 204 (1996) 5-8

A

B

Control

N2-hFGF Class I

7 mV

L

C

7 mV

I00 nn~

Lo,-

N2-bFGF Class 2

D

II0

I

130 pA l l0 pA 0pA -10 pA

pA

N 2 - h l q ; F (_'lass 3

_~

L__ IOOm.,,~

L

,

50pA

I 30pA 'l 10pA [-10pA /..'~0 pA

Fig. 3. Comparison of the firing patterns of El8 hippocampal neurons maintained in standard serum culture (control), or in N2-bFGF for 6 9 days. In 10 ng/ml bFGF, the ability to fire action potentials in response to depolarizing voltage steps varies with the neuronal class (see text). For class 1, n = 15, class 2, n = 13, class 3, n = 5, and standard, n = 20. Experimental details are described in the legend to Fig. 1.

Table 1). Although cla,;s 2 neurons can resemble those of class 1 (under DIC), when unambiguous identification was possible, we found they could more readily fire single, small, action potentials, but only on extreme depolarization (Fig. 3C, Table 1). Class 3 neurons, while distinct from those of classes 1 and 2, were harder to find owing to their scarcity (<5% of total MAP-2 positive neurons at day 7), and physical resemblance (under DIC) to

7

many of the glia present in the culture. However, the electrical characteristics of the class 3 neurons and glia were very different. Unlike glia, which did not fire action potentials, the class 3 neurons invariably fired multiple action potentials upon modest depolarization (Figure 3D). Electrophysiologically, class 3 neurons were essentially identical to their age-matched counterparts in serum culture (Fig. 3A), although the spike frequency was slightly lower. A limited analysis of individual currents also revealed differences among the three neuronal classes. All cells examined showed slowly inactivating K ÷ currents (Ik). However, greater differences were seen in the prevalence of Ca 2÷ currents between the classes (Table 1). These data clearly show that three neuronal classes exist in El8 bFGF cultures maintained for greater than 4 days. Such heterogeneity is at odds with recent claims that bFGF affords nearly pure populations of dividing neurons which differentiate [15], but our data are supported by immunocytochemical observations made on El6 cultures [18]. It is possible that the conflicting results reflect a marked sensitivity of neurons to bFGF culture conditions, especially passaging steps. Alternatively, the differentiated cells thought to originate from precursors dividing in culture [15] may in fact be class 3 neurons present at the time of plating which survive passaging. We recognize that MAP-2 staining and action potential generation may only provide a gross classification. Variability in the incidence of currents within each class certainly hints at further (electrophysiological) heterogeneity. As class 1 neurons are non-polarized, undergo mitosis, and have simple electrophysiology, they are almost certainly the proliferating precursors described by others [6,15]. In contrast, class 3 neurons seem to have escaped the mitogenic actions of bFGF [5,7]. Presumably, such cells either lack bFGF receptors, have altered second messenger responses to bFGF, or their bFGF receptors are desensitized [ 1,11 ]. Prolongation of action potential repolarization suggests class 3 neurons do respond to bFGF,

Table 1 Comparison of the firing properties and K + and Ca 2+ currents in El8 hippocampal neurons maintained in either standard serum culture or in N2-bFGF for 6-9 days Cell type

Class 1 Class 2 Class 3 Standard

Cells tiring action potentials

2 (15) 9 (13) 5 (5) 20 (20)

Vr (mV)

-67.3 -65.3 -66.7 -58.4

-+ 7.3 + 10.2 +_ 1.5 ± 1.9

Cells with K + currents

15 (15) 3 (3) ND 7 (7)

Cells with Ca 2+ currents

~

LVA

HVA

3 (15) 1 (3) ND 6 (7)

0 (3) 1 (7) 1 (1) 3 (3)

0 (3) 4 (7) 1 (1) 6 (7)

Total number of cells evaluated for each class is shown in parentheses. Current or voltage clamp recordings were made with an internal recording solution of K + gluconate (150 mM), KCI (5 raM), EGTA (1 mM), HEPES (10mM), and Mg-ATP (2mM), pH 7.2 adjusted with KOH, 275 _+5 mosmol. The external solution contained NaC1 (125 mM), KCI (2.5 mM), Na2HPO 4 (1.25 mM) MgCI 2 (2 mM), CaC12 (2 mM), NaHCO 3 (25 mM), and glucose (10 mM) at pH 7.4. Recording of K + current was done in the presence of 3/zM tetrodotoxin, and 30/~M TEA. CI was also included for Ca 2+ current studies. 1A, I k, LVA and HVA currents were classified according to their kinetics of inactivation and activation [9]. Neuronal input resistances (1228 +_328 Mfl) in N2-bFGF cultures were very similar to those for neurons in standard cultures (1104 + 242 Mr)).

8

J.H. Eubanks et al./Neuroscience Letters 204 (1996) 5-8

either directly or indirectly, through paths affecting K ÷ channels [13]. O f course, such b F G F signaling m a y arise indirectly through the glia we and others [18] find in b F G F cultures. Most puzzling are the class 2 neurons, as they have properties intermediate between class 1 and class 3 cells, and resemble j u v e n i l e (stage 3) control neurons [4]. It is possible that a linear transition exists where precursors divide and the j u v e n i l e progeny mature into adult neurons. However, this model does not explain our detection o f class 2 cell mitosis. A n alternative model, consistent with the failure of class 3 n e u r o n s to incorporate the cell division m a r k e r B r d U [8] in E l 8 b F G F / B r d U cultures, is that b F G F arrests the forward transition of precursors into neurons, but m a i n t a i n s or enhances the normal d e v e l o p m e n t and maturation of those cells which have already m a d e that transition. N e u r o n s plated into b F G F would thus be trapped in the states of differentiation found in the E l 8 hippocampus, and could yield considerable insight into hippocampal d e v e l o p m e n t a l ontogeny. W e would like to thank J. Stevens and J. Trogadis for excellent assistance with confocal microscopy and image reconstruction. This work was supported by grants to OTJ from the B l o o r v i e w Epilepsy Program, Sandoz Gerontological F o u n d a t i o n , Ontario Mental Heath Foundation, and the Medical Research C o u n c i l of Canada. [1] Baird, A., Fibroblast growth factors: activities and significance of non-neurotrophin neurotrophic growth factors, Curr. Opin. Neurobiol., 4 (1994) 78-86. [2] Bottenstein, J. Growth and differentiation of neural cells in defined media. In J. Bottenstein and G. Sato (Eds.), Cell Cultures in the Neurosciences, Plenum Press, New York, 1985, pp. 3-43. [3] Caceres, A., Banker, G. and Binder, L., lmmunocytoehemical localization of tubulin and microtubule-associated protein 2 during the development of hippocampal neurons in culture, J. Neurosci., 6 (1986) 714-722. [4] Dotti, C.G., The establishment of polarity by hippocampal neurons in culture, J. Neurosci., 8 (1988) 1454-1468.

[5] Eubanks, J.H., Kerr, R.G., Perez-Velasquez, J.L., Carlen, P.L., Mills, L.R. and Jones, O.T., Long-term bFGF neuronal culture: reintroduction into serum yields neurons and non-neuronal cells with neuronal characteristics, Neurosci. Lett., 194 (1995) 65-68. [6] Gage, F.H., Ray, J. and Fisher, L.J., Isolation, characterization, and use of stem cells from the CNS, Annu. Rev. Neurosci., 18 (1995) 159-192. [7] Goslin, K. and Banker, G., Rat hippocampal neurons in low density culture. In G. Banker and K. Goslin (Eds.), Culturing Nerve Cells, MIT, Cambridge, MA, 1991, pp. 207-226. [8] Gratzner, H.G., Monoclonal antibody to 5'-bromo- and 5'iododeoxyuridine: a new reagent for detection of DNA replication, Science, 218 (1982) 474-475. [9] Hille, B., Ionic Channels of Excitable Membranes (2nd edn.), Sinauer, Sunderland, MA, 1991. [10] Jacobson, M., Developmental Neurobiology (3rd edn.), Plenum Press, New York, 1991. [ll] Johnson, D. and Williams, L., Structural and functional diversity in the FGF receptor multigene family, Adv. Cancer Res., 60 (1993) 1-41. [12] Mayer, E., Dunnett, S.B. and Fawcett, J.W., Mitogenic effect of basic fibroblast growth factor on embryonic ventral mesencephalic dopaminergic neurone precursors, Dev. Brain Res., 72 (1993) 253-258. [13] Mills, L.R., Niessen, C.E., So, A.P., Carlen, P.L., Spigelman, I. and Jones, O.T., N-type Ca2+ channels are located on somata, dendrites, and a subpopulation of dendritic spines on live hippocampal pyramidal neurons, J. Neurosci., 14 (1994) 6815-6824. [14] Ray, J. and Gage, F.H., Spinal cord neuroblasts proliferate in response to basic fibroblast growth factor, J. Neurosci., 14 (1994) 3548-3564. [15] Ray, J., Peterson, D., Schinstine, M. and Gage, F.H., Proliferation, differentiation, and long-term culture of primary hippocampai neurons, Proc. Natl. Acad. Sci. USA, 90 (1993) 3602-3606. [16] Shaw, G., Banker, G. and Weber, K., An immunofluorescence study of neurofilament protein expressed by developing hippocampal neurons in tissue culture, Eur. J. Cell Biol., 39 (1985) 205-216. [17] Vescovi, A.L., Reynolds, B.A., Fraser, D.D. and Weiss, S., bFGF regulates the proliferative fate of unipotent (neuronal) and bipotent (neuronal/astroglial) EGF-generated CNS progenitor cells, Neuron, 11 (1993) 951-966. [18] Vicario-Abejon, C., Johe, J.K., Hazel, T.G., Collazo, D. and McKay, R.D.G., Functions of basic fibroblast growth factor and neurotrophins in the differentiation of hippocampal neurons, Neuron, 15 (1995) 105-114.