1641
Development 127, 1641-1649 (2000) Printed in Great Britain © The Company of Biologists Limited 2000 DEV9696
In vivo regulation of cell death by embryonic (pro)insulin and the insulin receptor during early retinal neurogenesis Begoña Díaz*, José Serna, Flora De Pablo and Enrique J. de la Rosa‡ Department of Cell and Developmental Biology, Centro de Investigaciones Biológicas, CSIC, Velázquez 144, E-28006 Madrid, Spain *Present address: Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720-3204, USA ‡Author for correspondence (e-mail:
[email protected])
Accepted 4 February; published on WWW 21 March 2000
SUMMARY Programmed cell death is an established developmental process in the nervous system. Whereas the regulation and the developmental role of neuronal cell death have been widely demonstrated, the relevance of cell death during early neurogenesis, the cells affected and the identity of regulatory local growth factors remain poorly characterized. We have previously described specific in vivo patterns of apoptosis during early retinal neurogenesis, and that exogenous insulin acts as survival factor (Díaz, B., Pimentel, B., De Pablo, F. and de la Rosa, E. J. (1999) Eur. J. Neurosci. 11, 1624-1632). Proinsulin mRNA was found to be expressed broadly in the early embryonic chick retina, and decreased later between days 6 and 8 of embryonic development, when there was increased expression of insulin-like growth factor I mRNA, absent or very scarce at earlier stages. Consequently, we studied whether proinsulin and/or insulin ((pro)insulin) action in prevention of cell death has physiological relevance during early neural development. In ovo
treatment at day 2 of embryonic development with specific antibodies against (pro)insulin or the insulin receptor induced apoptosis in the neuroretina. The distribution of apoptotic cells two days after the blockade was similar to naturally occurring cell death, as visualized by TdTmediated dUTP nick end labeling. The apoptosis induced by the insulin receptor blockade preferentially affected to the Islet1/2 positive cells, that is, the differentiated retinal ganglion cells. In parallel, the insulin survival effect on cultured retinas correlated with the activation of Akt to a greater extent than with the activation of MAP kinase. These results suggest that the physiological cell death occurring in early stages of retinal development is regulated by locally produced (pro)insulin through the activation of the Akt survival pathway.
INTRODUCTION
neuron cell death and the significance of cell death in early neural development has not been recognised. Although the pioneer work of Glücksmann (1951) described the incidence of cell death in the neuroepithelium of early embryos, this early neural apoptosis has been only recently characterized as affecting proliferating neuroepithelial cells and young differentiating neuroblasts (Homma et al., 1994; Blaschke et al., 1996, 1998; Weil et al., 1997; Díaz et al., 1999). An aspect even less defined than the process itself is the nature of the growth factors modulating cell survival/death during early neurogenesis of the central nervous system, at stages when local factors predominate over target-derived factors. Apoptosis has been described in the embryonic chick retina (Cuadros and Ríos, 1988; Martín-Partido, 1988; Frade et al., 1996a and 1997; Díaz et al., 1999), a classical vertebrate central nervous system model for the study of neurogenesis (Kahn, 1974; Rager, 1980; Prada et al., 1991). Its regulation, however, deserves a more detailed analysis. Exogenous insulin attenuates apoptotic cell death induced by growth factor deprivation in cultured neuroretinas and neurulating embryos (Díaz et al., 1999,
Programmed cell death is an important process in development (Jacobson et al., 1997; Vaux and Korsmeyer, 1999). There have been many observations of the process in the nervous system, including the initial description of physiological death (Vogt, 1842), the multiple examples presented in the seminal article of Glücksmann (1951) and, more recently, the inspiration for the generalization of the process (Raff et al., 1993). The neurotrophic theory, which establishes the role of targetderived neurotrophic factors in the protection of connecting neurons from programmed cell death, provided an essential conceptual basis for the study of cell death in the nervous system (Barde, 1989; Oppenheim, 1991). The increasing number of recent observations on the survival effect of other families of growth factors, derived from diverse sources, and affecting other neural cell types, has broadened the classical concept of neurotrophism (Korshing, 1993; Lewis et al., 1996; Voyvodic, 1996; Cameron et al., 1998; Pettmann and Henderson, 1998). Many of the investigations have centered on
Key words: Chick embryo, Neural development, Neuroepithelial cells, Ganglion cells, Apoptosis
1642 B. Díaz and others and Morales et al., 1997, respectively). The recognition of insulin as an embryonic growth/survival factor is still quite limited, in spite of the fact that it is expressed during early development in various vertebrate species (reviewed in De Pablo et al., 1990, and De Pablo and de la Rosa, 1995; Deltour et al., 1993), and that it is widely used to maintain healthy cell cultures (Barnes and Sato, 1980). Since insulin, largely in the form of unprocessed proinsulin, and insulin receptors are expressed during early chick retinal neurogenesis (de la Rosa et al., 1994; Hernández-Sánchez et al., 1995; Alarcón et al., 1998; García-de Lacoba et al., 1999), we examined the role of insulin in apoptosis by blocking (pro)insulin action by antibody treatment in ovo. This approach could also further clarify the relevance of (pro)insulin (a term used for convenience, to refer either to proinsulin or insulin when the assay does not discriminate between the primary product and the processed form) signaling for neural development. Additionally, it provides a more detailed characterization of the in vivo incidence of cell death during early neurogenesis, when most retinal cells are proliferating and the first type of neurons, the ganglion cells, are generating from those proliferating neuroepithelial cells. As hypothesized, both anti-insulin and anti-insulin receptor immunoglobulins (Igs) provoked an increase in apoptosis that preferentially affected young retinal ganglion cells. MATERIALS AND METHODS Chicken embryos White Leghorn embryos at the indicated ages were obtained by incubation of fertilized eggs (Granja Rodríguez-Serrano, Salamanca, Spain) at 38.4°C and 60-90% relative humidity. In situ hybridization In situ hybridization was performed with riboprobes on frozen tissue sections as described by de la Rosa et al. (1994), employing [α33P]UTP (Amersham Pharmacia Biotech, Rainham, Essex, UK) to increase the sensitivity, so that it was possible to characterize low mRNA levels not detected previously (de la Rosa et al., 1994). Hybridized sections were apposed either to Hyperfilm-βmax (Amersham) for 30 days or to a Phosphorimager screen (Molecular Dynamics, Sunnyvale, CA) for 7 days. The autoradiographic images were digitalized by scanning with a Ektron 1412 camera (Bedford, MA) and visualized with the NIH Image software. The control sense probe average value for each age was subtracted from the corresponding antisense probe digital image and the resulting image was represented in a pseudocolor scale. Antibodies and in ovo treatment Anti-insulin Igs were purified from an anti-porcine insulin antiserum raised in guinea pig (batch 627; Department of Pharmacology, Indiana University, Indianapolis, IN). Anti-IGF-I Igs were purified from an antihuman IGF-I antiserum raised in rabbit (kindly provided by Dr I. Torres-Alemán, Instituto Cajal, CSIC, Madrid, Spain). Anti-insulin receptor Igs were purified from the autoimmune antiserum of an insulinresistant patient (B-10; Van Obberghen et al., 1979). All these antisera recognize and block the corresponding chicken molecules (see Fig. 2C for Igs reactivity with the factors; De Pablo et al., 1982; Girbau et al., 1988; Fernández et al., 1998; and data not shown). Anti-insulin Igs recognize both purified chicken insulin (Litron Laboratory, Rochester, NY) and recombinant chicken proinsulin (cloned and expressed in E. coli in our laboratory; results unpublished). The endogenous product is therefore referred to as (pro)insulin. Anti-IGF-I Igs recognize chicken IGF-I (GroPep, Adelaide, Australia) and crossreact slightly with
proinsulin. Human IGF-II (GroPep) was not recognized by any of the purified Igs. Purification was performed from antisera, as well as from normal control sera (guinea pig, rabbit and human) by Affi-Gel Blue chromatography according to the manufacturer’s instructions (BioRad Laboratories, Hemel Hempstead, Hertfordshire, UK). After extensive dialysis against PBS, Igs purity and concentration were determined by SDS-PAGE and the Protein Colorimetric Assay (BioRad), respectively. Eggs were incubated for 50-60 hours prior to antibody treatment. A lateral window was cut in the shell, and the developmental stage of the embryo was assessed after vital staining with neutral red (Sigma, St. Louis, MO). Embryos of stages HH15-18 (Hamburger and Hamilton, 1951) were treated with 200 µg of anti-insulin Igs, 200 µg of anti-IGF-I Igs, 100 µg of anti-insulin receptor Igs or with the same amounts of the corresponding control Igs. The antibodies were applied over the embryo in a volume of 100 µl of PBS containing 0.7% (wt/vol) methylcellulose (Sigma). The shell was sealed with cellophane tape and the eggs were incubated for 48 hours. This is the time window described as being the most effective for the blocking effects of these antibodies in chick embryos (De Pablo et al., 1985). Some embryos received a second dose after 48 hours and were incubated for a further 48 hours. Specific embryonic death, macroscopic malformations or general growth delay were not observed during this period. Neuroretina organotypic culture Neuroretinas were isolated from chick embryos at day 4 of development (E4) and cultured in chemically defined medium as described by de la Rosa et al. (1998); Díaz et al. (1999). Where indicated, 10–7 M purified chicken insulin was added to the basal medium. Cultures were maintained for 6 hours at 37°C in a 5% CO2 atmosphere. This short culture period prevented conditioning of the medium, while it rendered significant differences in apoptotic cell death among treatments. Detection of apoptosis Apoptotic cell death was determined by two different methods. (1) Counting of pyknotic nuclei. The retinas were dissociated to obtain a single cell suspension, and the proportion of pyknotic nuclei was determined after nuclear staining with 4 µg/ml 4′,6-diamidino-2phenylindole (DAPI; Sigma) as described by de la Rosa et al. (1998); Díaz et al. (1999). (2) TdT-mediated dUTP nick end labeling (TUNEL) of the fragmented DNA was performed either in cryostat sections or in whole-mount retina, essentially as described by Morales et al. (1997); Díaz et al. (1999). Briefly, fixed tissues were permeated with 1% (w/v) Triton X-100, and digested with 20 µg/ml proteinase K (Boehringer Mannheim, Mannheim, Germany). When done in combination with immunostaining, proteinase K was replaced by 20 U/ml collagenase (Sigma). TUNEL reaction was performed with 0.1 U/ml terminal deoxynucleotidyl transferase and 10 µM biotin-16deoxy-UTP (Boehringer Mannheim) for 2.5 hours at 37°C, terminated by a 1-hour incubation at 65°C with 2 mM EDTA in PBS, and visualized by incubation with Cy2-streptavidin (Amersham). In the whole-mount retinas, pyknotic bodies were counted with a 40× objective (a field of 0.18 mm2) throughout the whole retina using a grid of 0.25 mm2 squares. Results are expressed by collating the squares with similar densities of pyknotic bodies to produce ‘isothanas’ (isodensity curves of dead cells, derived from the Greek thanatos, death; Díaz et al., 1999). The resulting isothanas were superimposed on the whole retina image. Images were captured with a Laser Confocal microscope (MCR 1024, BioRad). Immunostaining of neurons Differentiated neurons were identified in whole-mount retinas, retinal sections or dissociated retinal cells by staining with monoclonal antibodies against G4/Ng-CAM (1/1000 dilution; de la Rosa et al., 1990), and Islet1/2 (1/500; clone 39.4D5 from the Developmental Studies Hybridoma Bank; Austin et al., 1995). Islet1/2 staining was
Early neural cell death 1643
Fig. 1. Alternate proinsulin and IGF-I mRNA expression in early neurogenesis. Frozen head sections of E4 (A,E,I), E5 (B,F,J,M), E6 (C,G,K,N) and E8 (D,H,L) were hybridized with specific probes for proinsulin (A-D and M) or IGF-I (E-H and N) and exposed either to autoradiographic film (A-C, E-G, M and N) or to a Phosphorimager screen (D and H). The pseudocolor scale (A-H) reflects the highest signal in red and the lowest in blue. In M and N, the hybridization signals (in red) are superimposed over an image of the original tissue (dark field), also shown after staining with hematoxylin and eosin (I-L) to allow identification of structures. The main morphological features are indicated: III, third ventricle; cp, ciliary processes; d, diencephalon; eom, extraocular muscle; le, lens; m, mesencephalon; mt, metencephalon; nc, nasal cartilage; nr, neuroretina; op, olfactory process; pe, pigmented epithelium. Bar, 2 mm in A,E,I; 2.5 mm in B,F,J; 3.2 mm in C,G,K; 5 mm in D,H,L; 0.5 mm in M, and 0.8 mm in N. enhanced by microwave treatment for 10 minutes in 10 mM citrate buffer prior to incubation with the primary antibody. Staining was developed by consecutive incubations with biotin-conjugated goat anti-mouse Igs (1/200) and Cy2-streptavidin (1/200), or directly by incubation with Cy3-anti-mouse Igs (1/200; all from Amersham). Details of the different staining protocols can be found in de la Rosa et al. (1998). Images were obtained either with a confocal microscope, as above, or with a Zeiss Axioplan equipped with a cooled CCD camera Photometrics CH250/A (Tucson, AZ). Immunoblots For the determination of Akt, phospho-Akt, MAPK and phosphoMAPK protein levels, cultured E4 retinas were homogenized in RIPA buffer [10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% (wt/vol) Triton X-100 and protease inhibitor cocktail (Boehringer Mannheim)]. The extracts were fractionated in 10% SDS-PAGE and transferred to
nitrocellulose membranes in a semidry system using standard methods. Blots were incubated with one of the following primary antibodies: mouse monoclonal anti-MAP kinase (1/1000; Zymed Laboratories, San Francisco, CA), goat polyclonal anti-Akt1 (1/1000; Santa Cruz, Santa Cruz, CA), rabbit polyclonal anti-phosphoAkt (1/1000; New England BioLabs, Beverly, MA), and rabbit polyclonal anti-phospho-MAP kinase (1/1000; New England BioLabs). Blots were developed with the corresponding peroxidaseconjugated secondary antibodies (1/20000; Jackson Immunoresearch Laboratories, West Grove, PA) using the ECL system (Amersham). All blots were restained with mouse mAb anti-β-tubulin (1/10000; Sigma) to verify equal protein loading in all lanes. Phospho-Akt and phospho-MAP kinase levels were relative to total Akt-1 and MAP kinase levels developed in the same membrane. The determination of the anti-insulin and anti-IGF-I Igs specificity was performed by dot-blot. The different insulin family factors (0.2
1644 B. Díaz and others
A
pyknotic 2 bodies/mm 0-150 D 151-300 N T 301-500 V 501-750 >751
anti-Ins
anti-IR
control
control
B
untreated pyknotic 2 bodies/mm 0-250 D 251-750 N T 751-1500 V 1501-2500 >2501
basal
+ insulin Fig. 2. Incidence of apoptosis in E4 chick neuroretina. (A) E2 embryos were treated for 2 days with anti-insulin Igs (anti-Ins), antiIGF-I Igs, anti-insulin receptor Igs (anti-IR), the corresponding control Igs or remained untreated. The neuroretinas were dissociated to a single cell suspension and the DAPI-stained pyknotic nuclei were counted, out of a minimum of 1000 cells per experimental point. (B) E4 neuroretinas of untreated embryos were cultured either in basal medium or in the presence of 10-7 M chicken insulin (+Ins). After 6 hours in culture, the retinas were processed and counted as in A. Values in A and B represent the mean ± s.d. of the percentage of apoptotic cells found in the indicated number of single retinas (number inside the bar) in at least three independent experiments. Note the different y-axis scales. *, P
