Original Paper Acta Anat 1992;144:196-201
S.S.W. Tay W.C. Wong Department of Anatomy. Faculty of Medicine. National University of Singapore, Singapore
Key Words Insulin Immunoreactivity Monkey Spinal cord
Insulin-Like Immunoreactivity in the Monkey Spinal Cord
Abstract Insulin-like immunoreactive neurons were localized in the cervical, thoracic, lumbar and sacral segments of the monkey spinal cord. Both dorsal and ventral horn cells were labelled. Insulin-like reaction product was localized in the cell nucleus and cytoplasm. Both inner and outer nuclear membranes were labelled. Reaction product appeared to be scattered throughout the nucleoplasm but not within the nucleolus. In the cytoplasm, labelling was mainly localized in the cisternae of rER and saccules of Golgi apparatus. Both proximal and distal den drites were labelled, the reaction product was closely associated with the parallel arrays of neurotubules. Most of the distal dendrites were postsynaptic to nonlabelled axon terminals; however, some were postsynaptic to lightly labelled axon terminals. A labelled dendrite often formed the central element of a synap tic glomerulus with several nonlabelled axon terminals. It is hypothesized that insulin-like substance(s) may be modulating nuclear activities as well as neuro transmission at the synapse.
Introduction
Received: September 20. 1991 Accepted: February 28. 1992
[7] have also immunolocalized insulin in the neurons of the fetal rabbit brain. It is now well known that insulin may be synthesized in the CNS [6], So far. localization of insulin like immunoreactivity has been studied in the brain of vari ous mammals [8, 21]. Reddy et al. [22] have shown the presence of insulin-like immunoreactivity in the mamma lian spinal cord. However, little is known about the ultrastructural localization and possible function(s) of insulin in the spinal cord. The present study attempts to characterize the distribution of insulin-like immunoreactivity in the spinal cord of the monkey (Macaca fascicularis). A pre liminary report has been published [27].
S.S.W. Tay. PhD Department of Anatomy. Faculty of Medicine National University of Singapore 10 Kent Ridge Crescent, Singapore 0511 (Singapore)
© 1992 Kargcr AG. Basel 0001-5 180/92/1443-0196 $ 2.7570
Downloaded by: MacQuarie University 137.111.162.20 - 1/18/2020 10:34:57 AM
Insulin has traditionally been thought to have no effect on the mammalian brain. In recent years, several studies have demonstrated the presence of insulin and insulin receptors in the rat brain [14, 15, 23]. These studies have indicated that insulin may have a role in brain growth and maturation, neurotransmission and neuromodulation. The possible role of brain insulin in brain growth has been con firmed by specific insulin binding on fetal and neonatal brain membrane preparations [16, 17, 24, 30], The exact role of insulin in neurotransmission and neuromodulation remains unknown, although like other peptides in the CNS, insulin has been suggested to be involved in feedback loops in the rat brain [19], In cell cultures. Devaskar et al.
In the present study, a total of 6 adult monkeys of both sexes (body weight 2-2.5 kg) was used. For perfusion, each animal was anaesthe tized by an intraperitoneal injection of 0.5 ml of Sagatal (sodium pen tobarbital. 60 mg/ml) per kg body weight. Artificial respiration with air from a Harvard animal ventilator (model 683) through a tracheos tomy was done before thoracotomy. Five min before perfusion. 1.000 units of heparin and 2 ml of 1% sodium nitrite per kg body weight were given by intracardiac injection. The animals were rapidly perfused through the left cardiac ventri cle with 500 ml of Ringer's solution (pH 7.4) followed by 1.IKK) ml of fixative (4% paraformaldehyde+0.2% giutaraldehyde in 0.1 M phos phate buffer for electron microscopy and 4% paraformaldehyde alone in 0 .1 M phosphate buffer for light microscopy). The spinal cord was then dissected out and postfixed in the same fixative for an additional 4 h and kept overnight in 0.1 M phosphate buffer. After 2 washes at 15 min each, the spinal cord was sectioned with a razor blade and separated into cervical, thoracic, lumbar and sacral segments. Each segment was then cut with an Oxford vibratome. For both light and electron microscopy, 50-pm vibratome sections were cut and rinsed in phosphate-buffered saline (PBS). The free sections in vials were incubated in 4% normal goat serum (NGS) for 2 h at room temperature (20 °C). Sections were washed in PBS for I h with several changes and subsequently incubated in anti-insulin (Incstar Corp.. Minn., USA) diluted 1:500 with 1% NGS for 20-48 h at 4 °C. Subse quent antibody detection was carried out using the Vcctastain ABCKit (PK-4007. Vector Laboratories. Burlingame. Calif.. USA) against guinea pig IgG with 3.3’-diaminobenzidine as a peroxidase substrate, and intensified with nickel ammonium sulphate. The buffer solutions for L.M preparations contained 0.1% Triton X-100, whereas buffer solutions for EM preparations contained no Triton X-100. Sections for LM were mounted on gelatinized glass slides, dehydrated and coverslipped with Pcrmount. For electron microscopy, the freely floating sections were then osmicated and dehydrated in graded series of etha nol and embedded in Araldite mixture. Ultrathin sections were cut on a Reichert E ultramicrotome, stained with lead citrate only and viewed in JEOL 1200 CX or Philips 400T electron microscopes. Guinea-pig anti-insulin (Incstar Corp.) used at an antiserum dilu tion of 1:500 was satisfactory for both LM and EM preparations. Immunostaining was abolished by incubating the sections in 1% NGS minus the antiserum or by preincubation of the diluted antiserum with 50 gg/ml of synthetic insulin (Chemicon. Calif.. USA).
Results Light Microscopy Insulin-likc immunoreactive (ILI) neurons were local ized in the cervical, thoracic, lumbar and sacral segments of the spinal cord of all the monkeys studied. Immunolabelling was detected in neurons from both the dorsal and ven tral horns of the spinal cord. Staining appeared in the cell nucleus, cytoplasm as well as the dendritic processes of the neurons (fig. 1). Both large and small multipolar neurons were ILI, and not all labelled neurons and dendrites
showed the same staining intensity (fig.2, 3). A few small pyramidal-shaped cells also showed positive labelling for insulin-likc immunorcactivity (fig.2). Electron Microscopy The majority of the cell bodies were ovoid in shape, with a characteristic round nucleus situated either centrally or at the periphery. They showed a rich cytoplasm with a welldeveloped Golgi apparatus, cisternae of rER and numerous mitochondria. The insulin-like reaction product was immunolocalized in the cell nucleus (fig. 4-6) as well as in the cyto plasm (fig. 6). In the nucleus, immunolabelling was observed on the inner and outer nuclear membranes (fig.4,6); more over, the reaction product was also scattered throughout the matrix of the nucleoplasm (fig.4, 5). but not within the nucleolus (fig.5). In some cells, the nucleus (but not the nucleolus) was intensely stained (fig.4,5), while the nucleus in others was lightly stained (fig.6). In the cytoplasm. immu nolabelling for insulin-likc immunoreactivity was mainly localized in the cisternae of the rER (fig.6) as well as in the saccules of the Golgi apparatus. However, some cell cyto plasm appeared to be lightly labelled by the reaction, while others appeared to be nonlabelled (fig.4. 5). In the neuropil of the monkey spinal cord, the reaction product was almost exclusively localized in both proximal and distal dendrites (fig.7-9). The staining densities varied from one dendritic profile to another (cf. fig.7-9). A few weakly stained axon terminals were also detected in the neuropil (fig.7). In immunostained dendrites that were cut longitudinally, the reaction product appeared to be closely associated with parallel arrays of neurotubules. In cross sections of immunolabelled dendrites, the reaction product was clearly associated with the neurotubules and synaptic densities (fig.7, 9). Most of the ILI dendrites were postsynaptic to nonlabelled axon terminals containing numer ous small agranular vesicles and mitochondria (fig.9). However, a few labelled dendrites were postsynaptic to lightly labelled ILI axon terminals (fig.7). Moreover, a few close contacts involving immunolabelled dendrites were also observed. In other cases, a lightly stained ILI dendrite often formed the central element of a synaptic glomerulus with several nonlabelled axon terminals, which were char acterized by the presence of mitochondria and small agran ular vesicles in their axoplasm (fig.8).
197
Downloaded by: MacQuarie University 137.111.162.20 - 1/18/2020 10:34:57 AM
Materials and Methods
Fig. 1. Cross section of a monkey cervi cal cord. Labelled IL.I neurons arc found in both the dorsal and ventral horns. Note that the labelled neurons are all multipolar in nature. x32. Fig. 2. Higher magnification of the ven tral horn area of a monkey cervical cord. Note the darkly labelled multipolar ILI neu rons, with long dendritic processes. A few pyramidal cells are also labelled for insulinlike immunorcactivity (arrows). X87.5. Fig. 3. Higher magnification of the ven tral horn area of a monkey cervical cord. Observe the intense labelling of the multipo lar ILI neurons and their dendritic processes. A few ILI neurons appear to be lightly labelled for insulin-like immunoreactivity (arrows). Note also that the immunolabelling has masked the identification of the cell nucleus. X87.5.
Insulin has been identified in the brain of various mam malian species, including human [8,18], Various biological effects of insulin have been described earlier in the CNS, but the exact origin of the peptide remains unclear. Although several authors have hypothesized that the insu lin may be synthesized by the neurons themselves [3, 15], others have contended that insulin enters the brain via the
1%
TayAVong
blood vascular system [9]. Using radioactive labelled insu lin (i.e. I2>l-insulin), Frank et al. [9] have demonstrated that the peptide insulin is able to cross the neonatal rabbit blood-brain barrier, although the exact mechanism is still not clear. In another study. Baskin et al. [1] have also observed that insulin uptake of the adult rat hypothalamus is derived from the cerebrospinal fluid. They further showed that the cerebrospinal fluid insulin levels appear to correlate closely with the plasma insulin concentration. In
Insulin-Like Immunoreactivity in Spinal Cord
Downloaded by: MacQuarie University 137.111.162.20 - 1/18/2020 10:34:57 AM
Discussion
Fig. 4. Electron micrograph of a portion of an ILI neuron showing immunolabelling of the cell nucleus and the inner and outer nuclear membranes. Note that the cytoplasm of this neuron appears to be unlabelled. N = nucleus; rF.R = rough endoplasmic reti culum; GA = Golgi apparatus, x 13,400. Fig. 5. A portion of an immunolabclled neuron showing the scattered reaction prod uct in the nucleoplasm (arrows) and on the inner and outer nuclear membranes but not within the nucleolus. The cell cytoplasm appears to be unlabelled. N = nucleus; Nu = nucleolus. X 10,000. Fig. 6. A portion of an ILI neuron show ing immunolabelling in both the cell nucleus and cytoplasm. The reaction product is scat tered in the nucleoplasm (arrows), on the inner and outer nuclear membranes and within the rough endoplasmic reticulum. N = nucleus: rER = rough endoplasmic reti culum. x 10,(KM).
induce insulin action at intracellular locations. The inter nalization of insulin, as well as the existence of specific insu lin-binding sites at various intracellular structures and organelles, have been convincingly demonstrated by Gorden et al. [13]. In the present study, the localized insulin in the neurons of the spinal cord could have entered the neu rons by transcytosis or by the formation of receptor com plexes which are then internalized. Earlier studies have indicated that there is a predominant association of inter nalized insulin with lysosomes [4. 5], favouring the concept
199
Downloaded by: MacQuarie University 137.111.162.20 - 1/18/2020 10:34:57 AM
the present study, insulin-like immunoreactivity was found mainly in the nucleoplasm and nuclear membranes (but not the nucleolus), the cell cytoplasm and the dendrites. In the cytoplasm, immunolabelling was localized in the rER and Golgi apparatus. It is hypothesized that the insulin-like sub stance^) in the neuronal nucleus and cytoplasmic organ elles may have entered the neurons by crossing the bloodbrain barrier or via the cerebrospinal fluid. Some investigators have questioned the absolute need for a second messenger or transmitter system in order to
Fig. 7. Electron micrograph showing several labelled ILI dendrites (D) in the neu ropil of the monkey spinal cord. Most of these dendrites arc postsynaptic to nonlabelled axon terminals. The reaction product in dendrites appears to be closely associated with neurotubules and synaptic densities. At least two labelled dendrites are postsynaptic to lightly labelled axon terminals (AT,. AT,). Note that the reaction product in the labelled axon terminals is localized on the membranes of some of the synaptic vesicles (arrows). X 13,400. Fig. 8. A labelled ILI dendrite (D) forms the central element of a synaptic glomerulcs w'ith several axon terminals (AT), w'liich con tain mitochondria and numerous small agran ular vesicles. Observe the discrete reaction product which is closely associated with some neurotubules (arrows) of the dendrite, x lO.(XX). Fig. 9. A labelled ILI dendrite (D) is postsynaptic to a nonlabelled axon terminal (AT). Note the dense reaction product (arrows) associated with the neurotubules and synaptic densities, x 14.500.
2(X)
Tay/Wong
cell, such as the Golgi apparatus, rER and nuclear mem branes [2, 10-13, 20]. In view of the above findings, the present results from the monkey spinal cord indicate that insulin-like substance(s) may have been internalized via transcytosis or receptor-mediated transfer into the cell organelles, including the cell nucleus thereby modulating cellular activities at the cytoplasmic and nuclear levels. On the other hand, one should not preclude the possibil ity that the insulin-like substance(s) in the neurons of the monkey spinal cord may well be synthesized by the neurons
Insulin-Like Immunoreactivity in Spinal Cord
Downloaded by: MacQuarie University 137.111.162.20 - 1/18/2020 10:34:57 AM
that insulin is part of the degradation process. In later stud ies, internalized tracer insulin has been found to be associ ated with the rER and the cell nucleus [ 11,25], suggesting additional physiological significance of the internalization phenomenon of insulin. Thus, it has been proposed by vari ous workers that insulin via its internalization or its degra dation products [26] or the insulin receptor (28] might well be its own ‘second messenger'. This hypothesis may well be in accordance with the new finding that insulin-specific receptors have been localized at various sites within the
themselves. This possibility is borne out by the fact that insulin-like substance(s) appears to be localized in the rER and Golgi apparatus in the present as well as earlier studies [2. 12). More recent studies have also indicated that insulin mRNA is present in the brain, providing the essential bio chemical basis for the local synthesis of the peptide hor mone by the neuronal cells [3, 29], Further work involving a molecular approach is needed to clarify the exact origin of insulin-like substance(s) in the neurons of the mammalian spinal cord.
Acknowledgements The authors are grateful to Mr. Y.K. Han and Ms. Y.G. Chan for technical assistance. They would like to thank Mrs. C. Wong for typ ing the manuscript. This study was supported by grants RP 860343 and RP 870304 from the National University of Singapore.
References 9 Frank HJL. Jankovic-VokesT. Partridge WM. Morris WL: Enhanced insulin binding to blood-brain barrier in vivo and to brain microvessels in vitro in newborn rabbits. Diabetes 1985:34:728-733. 10 Goldfine ID: Insulin receptors and the site of action of insulin. Life Sci 1978:23:2639-2648. 11 Goldfine ID. Smith GJ: Binding of insulin to isolated nuclei. ProcNatl Acad Sei LJSA 1976: 73:1427-1431. 12 Goldfine ID. Smith GJ. Wong KY. Jones AL: Cellular uptake and nuclear binding of insulin in human cultured lymphocytes: Evidence for potential intracellular sites of insulin action. Proc Natl Acad Sci USA 1977:74:1368-1372. 13 Gorden P. Carpentier JL. Freychet P. Orci L: Internalization of polypeptide hormones. Mechanism, intracellular localization and sig nificance. Diabetologia 1980:18:263-274. 14 HavrankovaJ. RotliJ. Brownstcin MJ: Insulin receptors arc widely distributed in the central nervous system of the rat. Nature 1978:272: 827-829. 15 Havrankova J. Schmcchel D. Roth J. Brown stcin M.I: Identification of insulin in rat brain. Proc Natl Acad Sci USA 1978:75:5737-5741. 16 Kappy MS. Raizada MK: Adult-level insulin binding is present in term fetal rat CNS mem branes. Brain Res 1982:249:390-392. 17 Kappy M. Sellinger S. Raizada M: Insulin binding in four regions of the developing rat brain. J Neurochem 1984:42:198-203. 18 LeRoilh D. Hendricks SA. Lesniak MA. Rishi S. Becker KL. Havrankova J, Rosenzweig JL. Brownstcin MJ. Roth J: Insulin in brain and other extrapancreatic tissues of vertebrates and non-vertebrates. Adv Metab Disord 1983: 10:303-340. 19 Phillips Ml: Insulin in the brain: A feedback loop involving brain insulin and circumventricular organs: in Raizada. Phillips (eds): Insulin. Insulin-Like Growth Factors and Their Recep tors in the Central Nervous System. New York. Plenum Press. 1987. pp 163-175. 20 Posner BL Josefsberg Z. Bergeron JJM: Intra cellular polypeptide hormone receptors. Char acterization of insulin binding sites in Golgi fractions from the liver of female rats. J Biol Chem 1978:253:4067-4073.
21 Raizada MK: Localization of insulin-like immunoreactivity in the neurons from primary cultures of rat brain. Exp Cell Res 1983:143: 351-357. 22 Reddy NB. Engel WK. Askanas V: Different distributions of insulin-like immunoreactivity and insulin receptor in the mammalian spinal cord. Soc Neurosci Abstr 1987:13:1020. 73 Rosenzweig JL. HavrankovaJ. Lesniak MA. Brownstein MJ. Roth J: Insulin is ubiquitous in extrapancreatic tissues of rats and humans. Proc Natl Acad Sci USA 1980;77:572-576. 24 Sara VR. Hall K. Misaki M. Fryklund L. Chris tensen N. Wetterberg L: Ontogenesis of soma tomedin and insulin receptors in the human fetus. J Clin Invest 1983:71:1084-1094. 25 Solcr AP. Thomson KA. Smith RM. Jarett I : Immunological demonstration of the accumu lation of insulin, hut not insulin receptors, in nuclei of insulin-treated cells. Proc Natl Acad Sci USA 1989:86:6640-6644. 26 Steiner DF: Insulin today - The Banting Memorial Lecture. 1976. Diabetes 1976:26: 322-340. 27 Tay SSW. Wong WC: Ultrastructural localiza tion of insulin-like immunoreactivity in the monkey spinal cord. Anal Rec 1991:229: 88A-89A. 28 Van Obberghen E. Spooner PM. Kahn CR. Chernick SS. Garrison MM. Karlsson FA. Grunfeld C: Insulin-receptor antibodies mimic a late insulin effect. Nature 1979:280: 500-502. 29 Villa-Komaroff L. Gonzales A. Song HY. Wentworth B. Dobnes P: Novel insulin-related sequences in fetal brain. Adv F.xp Med Biol 1984:181:65-86. 30 Young WS III. Kuhar MJ. Roth J. Brownstcin MJ: Radiohistochemical localization of insulin receptors in the adult and developing rat brain. Neuropeptides 1980:1:15-22.
201
Downloaded by: MacQuarie University 137.111.162.20 - 1/18/2020 10:34:57 AM
1 Baskin DG. Woods SC. West DB. Van llouten M. Posner Bl. Dorsa DM. Porte D Jr: Immunocytochemical detection of insulin in rat hypothalamus and its possible uptake from cerebrospinal fluid. Endocrinology 1983:113: 1818-1825. 2 Bergeron JJM. Posner Bl. Josefsberg Z. Sikstrom R: Intracellular polypeptide hormone receptors. The demonstration of specific bind ing sites for insulin and human growth hor mone in Golgi fractions isolated from the liver of female rats. J Biol Client 1978:253: 4058-4066. 3 Boyd F r Jr. Clarke DW. M utherTF. Raizada MK: Insulin receptors and insulin modulation of norepinephrine uptake in neuronal cultures from rat brain. J Biol Chem 1985:260: 15880-15884. 4 Carpentier JL. Gorden P. Amherdl M. Van Obberghen E. Kahn CR. Orci L: '"'l-insulin binding to cultured human lymphocytes. Initial localization and fate of hormone determined by quantitative electron microscopic autoradi ography. J Clin Invest 1978:61:1057-1070. 5 Carpentier JL. Gorden P. Freychct P. l.e Cam A. Orci L: Binding, internalization and lyso somal association of '"T-insulin in isolated rat hepatocytes. Expericntia 1978:34:936. 6 Clarke DW. PoulakosJJ. Mudd I.M. Raizada MK. Cooper DL: Evidence for central nervous system insulin synthesis: in Raizada. Phillips (eds): Insulin. Insulin-Like Growth Factors and Their Receptors in the Central Nervous System. New York, Plenum Press. 1987. pp 121-130. 7 Devaskar S. Schechter R. Kahn A: Localiza tion of insulin to neuronal cells: in Raizada. Phillips (eds): Insulin. Insulin-Like Growth Factors and Their Receptors in the Central Nervous System. New York. Plenum Press. 1987. pp 131-137. 8 Dorn A. Bernstein MG. Rinnc A. Malin ll-J. Ziegler M: Insulin-like immunoreactivity in the human brain. Histochemistry 1982:74: 293-300.