- Email: [email protected]
Perivascular microglia in the rat neural lobe engulf magnocellular secretory terminals during osmotic stimulation
Neuroscience Letters 180 (1994) 235-238
ELSEVIER
HEUROSCItNCt LETTERS
Perivascular microglia in the rat neural lobe engulf magnocellular secretory terminals during osmotic stimulation Thomas H. Mander**, John F. Morris* Department of Human Anatomy, University of Oxford, South Parks Road, Oxford 0)(1 3QX, UK Received 2 June 1994; Revised version received 2 September 1994; Accepted 6 September 1994
Abstract The response of microglia in the rat neural lobe to osmotic stimulation has been studied. Microglia were identifiedby immunoreactivity for the macrophage markers OX-42 and F4/80. The numerical density of microglia did not change significantlywith osmotic stimulation but microglia in the perivascular space partially or completelyenclosed significantlygreater numbers of neurosecretory terminals in osmotically stimulated animals. Key words." Microglia; Neurohypophysis; Osmotic stress; Neurosecretory terminal; Rat, Long-Evans
The hypothalamo-neurohypophysial system (HNS) consists of magnocellular neurons with cell bodies in the hypothalamus and axons projecting to the neural lobe. The neurons are closely associated with two types of glial cells, astrocytes/pituicytes and microglia, myelomonocytic cells of bone marrow origin. Activation of the HNS by osmotic stimulation causes a decrease in the extent to which magnocellular neurosecretory cell bodies and axons are enclosed by astrocyte/ pituicyte processes and an increase in the extent to which the axonal terminals cover the perineuronal basal lamina [4,19,20,22]. However, little is known about the response of microglia in the neural lobe to osmotic stress. In the neural lobe, cells resembling microglia that contact and occasionally phagocytose neurosecretory axons protruding beyond the perineuronal basal lamina have been described [13] and their identity as macrophages demonstrated immunocytochemically [16]. P~rez-Figares et al. [14] reported increased numbers of microglia in the neural lobe of rats given 2% saline to drink for 7 days but gave no quantitative data. We have performed a quanti*Corresponding author. Fax: (44) (865) 272164. E-mail: [email protected].
**Present address." Xenova Ltd, 240 Bath Road, Slough, Berkshire SLI 4EF, UK. 0304-3940/94/$7.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0304-3940(94)00690-3
tative study using macrophage-directed antibodies to identify the microglia and electron microscopy to examine the association between microglia and neurosecretory terminals during and after osmotic stress. Adult male Long-Evans rats, aged 3 - 6 months and weighing 300- 450 g, were placed in single metabolic cages at constant room temperature and humidity, with lights on from 06: 00 to 20: 00. The rats were divided into three groups given: tap water for 72 h ('water' group); 2% saline for 72 h ('saline' group); or 2% saline for 72 h and then tap water ad libitum for 48 h ('recovery' group). For light microscopic immunocytochemistry, animals were anaesthetized and perfused via the heart with saline and periodate-lysine-paraformaldehyde fixative at pH 7.0 [12]. Pituitaries were isolated and immersion fixed for 4 h at 4°C, infiltrated overnight with 10% sucrose, surrounded in Tissue-Tek, frozen and transverse 10-/1mthick cryostat sections prepared. Sections were blocked with 0.25% BSA, then incubated with either a 1:1000 dilution of rabbit polyclonal antiserum raised against partially purified murine F4/80 (a gift from Professor S. Gordon, Sir William Dunn School of Pathology, University of Oxford, Oxford, UK), a macrophage-specific 160 kDa plasma membrane glycoprotein [18], or mouse monoclonal OX-42 diluted 1:250 (Serotec, Kidlington, UK) that binds to the complement type 3 receptor [17].
236
7~H. M a n d e z J. E Morris/Neuroscience Let ters 180 ,"1994) 235 23,¥
Antibody-binding sites were localized by use of the ABC method (Vector laboratories, Peterborough, UK) and visualized by DAB substrate (Sigma, Poole, UK). Parallel control sections processed with pre-immune serum at the same dilution showed no staining. For ultrastructural studies, anaesthetized rats were perfused via the heart with 2.5% glutaraldehyde and 4% formaldehyde in 0.1 M phosphate buffer, pH 7.3. Neural lobes were isolated, fixed overnight at 4°C, immersed in 1% osmium tetroxide in 0.1 M phosphate buffer for 60 min, washed, dehydrated and embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined by use of a JEOL transmission electron microscope operated at 80 kV. The immunocytochemically stained slides were used visually to assess the number of macrophages and their immunophenotype. Using the ultrathin sections, the number of transnuclear sectioned profiles of perivascular and interstitial microglia (defined ultrastructurally) in each complete grid square of tissue (average 14 squares/ animal) was counted and the number of cells/l 03/~m 2 was calculated. Micrographs of transnuclear profiles of perivascular microglia selected by a systematic, random procedure from non-adjacent grid squares were prepared and, from each animal, the outlines of 10 cells was drawn at x 8850 final magnification. The area, perimeter, number and length of processes (total distance along process from the cell body) and length of plasmalemma in direct contact with other cellular components were determined for each cell, together with the number of magnocellular neurosecretory terminals either partially or completely enclosed by microglial processes in the plane of section. Mean data from each animal were examined by the Oxstat statistical package for significant (P < 0.05) differences by ANOVA and unpaired Student's t tests. Data are expressed as mean + S.E.M. values for n = 3 animals/group. Microglia labelled by F4/80 and OX-42 antibodies were prominent in all groups of animals (Fig. 1); no obvious changes in their number or immunophenotype emerged. Ultrastructurally, microglia had elongated nuclei with dense peripheral heterochromatin, a narrow rim of electron-dense cytoplasm containing prominent rough endoplasmic reticulum, lysosomes, vacuoles, mitochondria and dense bodies (Fig. 2). Microglia were present in perivascular spaces and interstitium, with ~75% in the perivascular spaces. There was no significant difference in the numerical density (cell profiles/103/lm 2 section) of either perivascular or interstitial microglia between the three groups (perivascular: water- 4.0 _+ 0.2: saline 3.8 + 0.4; recovery - 4.9 + 0.6; interstitial microglia: water- 1.5 + 0.1; saline-2.0 + 0.6; recovery 1.5 + 0.2; total: w a t e r - 5.5 + 0.2; saline 5.8 + 0.8, recovery 6.4 + 0.4). Morphometric analysis of the perivascular microglia revealed that mean values for total area, nuclear area, number of processes per cell and process
Fig. 1. Light micrographs of longitudinal sections through the neural lobe of rats given (a) water or (b) 2% saline to drink for 3 days. Microglia immunostained for OX-42 with irregular morphology and processes arc randomly scattered across the section at similar numerical density, by, blood vessel. Bar = 20 lira.
lengths were highest in the saline group (not shown) but only the perimeters were significantly increased (water 27 _+ 2/Ira; saline - 35 + 1/~m, P < 0.05). Data for the recovery group were generally intermediate between the water and saline groups. Between 20 25% of the total plasmalemma of perivascular microglia in all three groups was in direct contact with other cellular components of the perivascular space. These were most frequently neurosecretory nerve terminals projecting beyond the basal lamina. The terminals were either partly or completely enclosed by microglial processes (Fig. 2a,b). The number of terminals enclosed was significantly increased in the saline group (water 5.0 + 1.2; saline 9.3 _+ 0.3 terminals/10 microglial cells; P < 0.05). Interestingly, only perivascular microglia from the saline group enclosed more than one terminal (Fig. 2b). Microglia form -20% of the glial cells in the rat neural lobe, the other cells being pituicytes [16]. It is well documented that, when vasopressin and oxytocin secretion is high during dehydration and lactation, pituicytes enclose fewer neurosecretory terminals [7]. Our analysis shows that the relationship between microglia and neurosecretory terminals also alters during osmotic stimulation. The number of microglia in the rat neural lobe may increase after 7 days salt loading [14]. Microglia in the mouse neural lobe synthesize DNA in response to osmotic stimulation, peaking at 48 h [9] and our own experiments using double labelling for bromodeoxyuridineand macrophage-specific (F4/80) antibodies (T.H. Mander and J.F. Morris, unpubl, data) support this finding. Although we could not demonstrate a significant in-
T.H. Mander. J.E Morris/Neuroscience Letters 180 (1994) 235-238
237
Fig. 2. Electron micrographs of perivascular microglia in the neural lobe of adult rats having drunk (a) water and (b) 2% saline for 3 days. ~, perineuronal basal lamina, a: a typical microglial cell with dense nuclear chromatin and prominent cytoplasmic organelles partially encloses a neurosecretory terminal (nl). Another microglial process is in direct contact (fine arrows) with two other terminals (n2, n3). b: the larger microglial cell profile has numerous processes containing prominent rough endoplasmic reticulum which are engulfing several neurosecretory terminals two of which (*) are completely enclosed, bv, blood vessel. Insert detail. Bars = 1 ~um.
crease in the numerical density of microglia after 3 days saline drinking the total number of microglia may well have increased because the neural lobe increases in size during salt loading [6]. Taken together, the data suggest that microglia divide to maintain a constant numerical density in an enlarging neural lobe. During osmotic stimulation, pituicytes rapidly retract their processes so that a greater proportion of the perineuronal basal lamina is covered by neurosecretory terminals [2,15,19]. The changes are reversed on rehydration [23] and this plasticity also occurs in pregnancy, parturition and lactation [21]. The mechanisms control-
ling the changes appear to involve fl-adrenergic receptors [11] but vasopressin V1 receptors on pituicytes [8] may also be involved. In contrast, the microglia in the perivascular spaces of rat neural lobes envelop more rather than fewer neurosecretory terminals in osmotically stimulated animals compared with normally hydrated controls and this is reversed after 48 h rehydration. The increased enclosure of neurosecretory terminals during osmotic stimulation probably precedes phagocytosis of the terminals, which occurs also in normal rats [16]. This may be because more neurosecretory neurons protrude into the perivas-
238
T H. Mandep: .L E Morris/Neuroscience Letters 180 (1994) 235-238
cular spaces after retraction of the pituicyte processes. Oxytocinergic neurons become more closely associated with hypothalamic blood vessels in late pregnancy/early parturition [3], so magnocellular neurons are able to extend their processes in response to altered demand. In addition, the perivascular space may be reduced because stimulation with high K + reduces the perivascular space in isolated neural lobes [1] and secretory activity elevates extracellular K + [10]. Neural lobe microglia surround and phagocytose neurosecretory nerve terminals, particularly when these extend into the perivascular space, in normal animals [16] and to an increased extent on osmotic stimulation. Neurosecretory neurons project to a basal lamina rather than to another neuron and can regenerate to form a miniature neural lobe after hypophysectomy [5]. Thus, it seems likely that neurosecretory axons continually grow and that microglia recognize terminals that obtrude through the perineuronal basal lamina and remove them to maintain a clear perivascular space. This axonal growth may increase at the same time as the cells hypertrophy during increased magnocellular neurosecretory activity. What signal causes the microglia to recognize the protruding terminals and to phagocytose them remains to be discovered. It will be interesting to discover whether the perivascular macrophages show similar changes in chronic salt loading, pregnancy/parturition and lactation. This study was supported by an MRC studentship and Oxford University. We thank D. Hardiman for expert animal care, A. Loreto-Gardner for typing the manuscript, and L. Scott and S. Rodgers for technical assistance. Professor S. Gordon, and Drs. V.H. Perry and L.J. Lawson provided antibodies and helpful advice. Ill Armstrong~ W.E., Tian, M. and Reger, J.F., Elevated extracellular potassium is associated with a reduced extracellular space in rat neural lobe in vitro, J. Neurocytol., 20 (1991) 564 572. [2] Beagley, G.H. and Hatton, G.I., Rapid morphological changes in supraoptic nucleus and posterior pituitary induced by a single hypertonic saline injection, Brain Res. Bull., 28 (1992) 613 618. [3] Blanco, E., Pilgrim, C., Vazquez, R. and Jirikowski, G.F., Plasticity of the interface between oxytocm neurons and the vasculature in late pregnancy: an ultrastructural morphometric study, Acta Histochem.,91 (1991) 165 172. [4] Chapman, D.B., Theodosis, D.T., Montagnese, C., Poulain. D.A. and Morris, J.F., Osmotic stimulation causes structural plasticity of neurone glia relationships of the oxytocin but not vasopressin secreting neurones in the hypothalamic supraoptic nucleus, Neuroscience. 17 (1986) 679 686. [5] Dellmann, H.-D., Stoeckel, M.E., Porte, A. and Stutinsky, F., Ultrastructure of the neurohypophysial glial cells following stalk transection in the rat, Experientia, 30 (1974) 1220 1222.
[6] Duchen, L.W., Changes in the volume of the lobes of the pituitary gland and in the weight and water content of organs of rats given hypertonic saline, J. Endocrinol., 41 (1968) 593-600. [7] Hatton, G.I., Pituicytes, glia and control of terminal secretion, J. Exp. Biol., 139 (1988) 67-79. [8] Hatton, G.I., Bicknell, R.J., Hoyland, J., Bunting, R. and Mason, W.T., Arginine vasopressin mobilises intracellular calcium via V I receptor activation in astrocytes (pituicytes) cultured from adult rat neural lobes, Brain Res., 588 (1992) 75 83. [9] Lawson, L.J., Perry, V.H. and Gordon, S., Microglial responses to physiological change: elevation of DNA synthesis without significant activation in the neurohypophysis of the osmotically stressed mouse, Neuroscience, 56 (1993) 929 938. [10] Leng, G. and Shibuki, K., Extracellular potassium changes in the rat neurohypophysis during activation of the magnocellular neurosecretory system, J. Physiol., 392 (1987) 97-112. [11] Luckman, S.M. and Bicknell, R.J., Morphological plasticity that occurs in the neurohypophysis following activation of the magnocellular neurosecretory system can be mimicked in vitro by fl-adrenergic stimulation, Neuroscience, 39 (1990) 701 709. [12] McLean, I.W. and Nakane, RK., Periodate-lysine-paraformaldehyde fixative: a new fixative for immuno-eleetron microscopy, ,1. Histochem. Cytochem., 22 (1974) 1077-I083. [13] Olivieri-Sangiacomo, C., On the fine structure of the perivascular cells in the neural lobe of rats, Z. Zellforsch., 132 (1972) 25 34. [14] P6rez-Figares, J.M., Fernandez-Llebrez, R, Pdrez, J. and MarinGir6n, F., Presence of microglia-like cells in the neurohypophysis of normal and dehydrated rats, Acta Anat., 127 (1986) 195 200. [15] Perlmutter, L.S., Hatton, G.I. and Tweedle, C.D., Plasticity in the in vitro neurohypophysis: effects of osmotic changes on pituicytes, Neuroscience, 12 (1984) 503 511. [16] Pow, D.V., Perry, V.H., Morris, J.F. and Gordon, S., Microglia in the neurohypophysis associate with and endocytose terminal portions of neurosecretory neurons, Neuroscience, 33 (1989) 567 578. [17] Robinson, A.R, White, T.M. and Mason, D.W., Macrophage heterogeneity in the rat as delineated by two monoclonal antibodies MRC OX-41 and MRC OX-42, the latter recognizing complement receptor type 3, Immunology, 57 (1986) 239 247. [18] Starkey, RM., Turley, L. and Gordon, S., The mouse macrophage-specific glycoprotein defined by monoclonal antibody F4/ 80: characterization, biosynthesis and demonstration of a rat analogue, Immunology, 60 (1987) 117122. [19] Tweedle, C.D. and Hatton, G.I., Glial cell enclosure of neurosecretory endings in the neurohypophysis of the rat, Brain Res., 192 (1980) 555 559. [20] Tweedle, C D . and Hatton, G.I., Evidence for dynamic interactions between pituicytes and neurosecretory axons in the rat, Neuroscience, 5 (1980) 66t 667. [21] Tweedle, C.D. and Hatton, G.I., Magnocellular neuropeptidergic terminals in neurohypophysis: rapid glial release of enclosed axons during parturition, Brain Res. Bull., 8 (1982) 205 209. [22] Tweedle, C.D. and Hatton, G.I., Synapse formation and disappearance in adult rat supraoptic nucleus during different hydration states, Brain Res., 309 (1984) 373 376. [23] Tweedle, C.D. and Hatton, G.I., Morphological adaptability at neurosecretory axonal endings on the neurovascular contact zone of the rat neurohypophysis, Neuroscience, 20 (1987) 241 246.
ELSEVIER
HEUROSCItNCt LETTERS
Perivascular microglia in the rat neural lobe engulf magnocellular secretory terminals during osmotic stimulation Thomas H. Mander**, John F. Morris* Department of Human Anatomy, University of Oxford, South Parks Road, Oxford 0)(1 3QX, UK Received 2 June 1994; Revised version received 2 September 1994; Accepted 6 September 1994
Abstract The response of microglia in the rat neural lobe to osmotic stimulation has been studied. Microglia were identifiedby immunoreactivity for the macrophage markers OX-42 and F4/80. The numerical density of microglia did not change significantlywith osmotic stimulation but microglia in the perivascular space partially or completelyenclosed significantlygreater numbers of neurosecretory terminals in osmotically stimulated animals. Key words." Microglia; Neurohypophysis; Osmotic stress; Neurosecretory terminal; Rat, Long-Evans
The hypothalamo-neurohypophysial system (HNS) consists of magnocellular neurons with cell bodies in the hypothalamus and axons projecting to the neural lobe. The neurons are closely associated with two types of glial cells, astrocytes/pituicytes and microglia, myelomonocytic cells of bone marrow origin. Activation of the HNS by osmotic stimulation causes a decrease in the extent to which magnocellular neurosecretory cell bodies and axons are enclosed by astrocyte/ pituicyte processes and an increase in the extent to which the axonal terminals cover the perineuronal basal lamina [4,19,20,22]. However, little is known about the response of microglia in the neural lobe to osmotic stress. In the neural lobe, cells resembling microglia that contact and occasionally phagocytose neurosecretory axons protruding beyond the perineuronal basal lamina have been described [13] and their identity as macrophages demonstrated immunocytochemically [16]. P~rez-Figares et al. [14] reported increased numbers of microglia in the neural lobe of rats given 2% saline to drink for 7 days but gave no quantitative data. We have performed a quanti*Corresponding author. Fax: (44) (865) 272164. E-mail: [email protected].
**Present address." Xenova Ltd, 240 Bath Road, Slough, Berkshire SLI 4EF, UK. 0304-3940/94/$7.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0304-3940(94)00690-3
tative study using macrophage-directed antibodies to identify the microglia and electron microscopy to examine the association between microglia and neurosecretory terminals during and after osmotic stress. Adult male Long-Evans rats, aged 3 - 6 months and weighing 300- 450 g, were placed in single metabolic cages at constant room temperature and humidity, with lights on from 06: 00 to 20: 00. The rats were divided into three groups given: tap water for 72 h ('water' group); 2% saline for 72 h ('saline' group); or 2% saline for 72 h and then tap water ad libitum for 48 h ('recovery' group). For light microscopic immunocytochemistry, animals were anaesthetized and perfused via the heart with saline and periodate-lysine-paraformaldehyde fixative at pH 7.0 [12]. Pituitaries were isolated and immersion fixed for 4 h at 4°C, infiltrated overnight with 10% sucrose, surrounded in Tissue-Tek, frozen and transverse 10-/1mthick cryostat sections prepared. Sections were blocked with 0.25% BSA, then incubated with either a 1:1000 dilution of rabbit polyclonal antiserum raised against partially purified murine F4/80 (a gift from Professor S. Gordon, Sir William Dunn School of Pathology, University of Oxford, Oxford, UK), a macrophage-specific 160 kDa plasma membrane glycoprotein [18], or mouse monoclonal OX-42 diluted 1:250 (Serotec, Kidlington, UK) that binds to the complement type 3 receptor [17].
236
7~H. M a n d e z J. E Morris/Neuroscience Let ters 180 ,"1994) 235 23,¥
Antibody-binding sites were localized by use of the ABC method (Vector laboratories, Peterborough, UK) and visualized by DAB substrate (Sigma, Poole, UK). Parallel control sections processed with pre-immune serum at the same dilution showed no staining. For ultrastructural studies, anaesthetized rats were perfused via the heart with 2.5% glutaraldehyde and 4% formaldehyde in 0.1 M phosphate buffer, pH 7.3. Neural lobes were isolated, fixed overnight at 4°C, immersed in 1% osmium tetroxide in 0.1 M phosphate buffer for 60 min, washed, dehydrated and embedded in epoxy resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined by use of a JEOL transmission electron microscope operated at 80 kV. The immunocytochemically stained slides were used visually to assess the number of macrophages and their immunophenotype. Using the ultrathin sections, the number of transnuclear sectioned profiles of perivascular and interstitial microglia (defined ultrastructurally) in each complete grid square of tissue (average 14 squares/ animal) was counted and the number of cells/l 03/~m 2 was calculated. Micrographs of transnuclear profiles of perivascular microglia selected by a systematic, random procedure from non-adjacent grid squares were prepared and, from each animal, the outlines of 10 cells was drawn at x 8850 final magnification. The area, perimeter, number and length of processes (total distance along process from the cell body) and length of plasmalemma in direct contact with other cellular components were determined for each cell, together with the number of magnocellular neurosecretory terminals either partially or completely enclosed by microglial processes in the plane of section. Mean data from each animal were examined by the Oxstat statistical package for significant (P < 0.05) differences by ANOVA and unpaired Student's t tests. Data are expressed as mean + S.E.M. values for n = 3 animals/group. Microglia labelled by F4/80 and OX-42 antibodies were prominent in all groups of animals (Fig. 1); no obvious changes in their number or immunophenotype emerged. Ultrastructurally, microglia had elongated nuclei with dense peripheral heterochromatin, a narrow rim of electron-dense cytoplasm containing prominent rough endoplasmic reticulum, lysosomes, vacuoles, mitochondria and dense bodies (Fig. 2). Microglia were present in perivascular spaces and interstitium, with ~75% in the perivascular spaces. There was no significant difference in the numerical density (cell profiles/103/lm 2 section) of either perivascular or interstitial microglia between the three groups (perivascular: water- 4.0 _+ 0.2: saline 3.8 + 0.4; recovery - 4.9 + 0.6; interstitial microglia: water- 1.5 + 0.1; saline-2.0 + 0.6; recovery 1.5 + 0.2; total: w a t e r - 5.5 + 0.2; saline 5.8 + 0.8, recovery 6.4 + 0.4). Morphometric analysis of the perivascular microglia revealed that mean values for total area, nuclear area, number of processes per cell and process
Fig. 1. Light micrographs of longitudinal sections through the neural lobe of rats given (a) water or (b) 2% saline to drink for 3 days. Microglia immunostained for OX-42 with irregular morphology and processes arc randomly scattered across the section at similar numerical density, by, blood vessel. Bar = 20 lira.
lengths were highest in the saline group (not shown) but only the perimeters were significantly increased (water 27 _+ 2/Ira; saline - 35 + 1/~m, P < 0.05). Data for the recovery group were generally intermediate between the water and saline groups. Between 20 25% of the total plasmalemma of perivascular microglia in all three groups was in direct contact with other cellular components of the perivascular space. These were most frequently neurosecretory nerve terminals projecting beyond the basal lamina. The terminals were either partly or completely enclosed by microglial processes (Fig. 2a,b). The number of terminals enclosed was significantly increased in the saline group (water 5.0 + 1.2; saline 9.3 _+ 0.3 terminals/10 microglial cells; P < 0.05). Interestingly, only perivascular microglia from the saline group enclosed more than one terminal (Fig. 2b). Microglia form -20% of the glial cells in the rat neural lobe, the other cells being pituicytes [16]. It is well documented that, when vasopressin and oxytocin secretion is high during dehydration and lactation, pituicytes enclose fewer neurosecretory terminals [7]. Our analysis shows that the relationship between microglia and neurosecretory terminals also alters during osmotic stimulation. The number of microglia in the rat neural lobe may increase after 7 days salt loading [14]. Microglia in the mouse neural lobe synthesize DNA in response to osmotic stimulation, peaking at 48 h [9] and our own experiments using double labelling for bromodeoxyuridineand macrophage-specific (F4/80) antibodies (T.H. Mander and J.F. Morris, unpubl, data) support this finding. Although we could not demonstrate a significant in-
T.H. Mander. J.E Morris/Neuroscience Letters 180 (1994) 235-238
237
Fig. 2. Electron micrographs of perivascular microglia in the neural lobe of adult rats having drunk (a) water and (b) 2% saline for 3 days. ~, perineuronal basal lamina, a: a typical microglial cell with dense nuclear chromatin and prominent cytoplasmic organelles partially encloses a neurosecretory terminal (nl). Another microglial process is in direct contact (fine arrows) with two other terminals (n2, n3). b: the larger microglial cell profile has numerous processes containing prominent rough endoplasmic reticulum which are engulfing several neurosecretory terminals two of which (*) are completely enclosed, bv, blood vessel. Insert detail. Bars = 1 ~um.
crease in the numerical density of microglia after 3 days saline drinking the total number of microglia may well have increased because the neural lobe increases in size during salt loading [6]. Taken together, the data suggest that microglia divide to maintain a constant numerical density in an enlarging neural lobe. During osmotic stimulation, pituicytes rapidly retract their processes so that a greater proportion of the perineuronal basal lamina is covered by neurosecretory terminals [2,15,19]. The changes are reversed on rehydration [23] and this plasticity also occurs in pregnancy, parturition and lactation [21]. The mechanisms control-
ling the changes appear to involve fl-adrenergic receptors [11] but vasopressin V1 receptors on pituicytes [8] may also be involved. In contrast, the microglia in the perivascular spaces of rat neural lobes envelop more rather than fewer neurosecretory terminals in osmotically stimulated animals compared with normally hydrated controls and this is reversed after 48 h rehydration. The increased enclosure of neurosecretory terminals during osmotic stimulation probably precedes phagocytosis of the terminals, which occurs also in normal rats [16]. This may be because more neurosecretory neurons protrude into the perivas-
238
T H. Mandep: .L E Morris/Neuroscience Letters 180 (1994) 235-238
cular spaces after retraction of the pituicyte processes. Oxytocinergic neurons become more closely associated with hypothalamic blood vessels in late pregnancy/early parturition [3], so magnocellular neurons are able to extend their processes in response to altered demand. In addition, the perivascular space may be reduced because stimulation with high K + reduces the perivascular space in isolated neural lobes [1] and secretory activity elevates extracellular K + [10]. Neural lobe microglia surround and phagocytose neurosecretory nerve terminals, particularly when these extend into the perivascular space, in normal animals [16] and to an increased extent on osmotic stimulation. Neurosecretory neurons project to a basal lamina rather than to another neuron and can regenerate to form a miniature neural lobe after hypophysectomy [5]. Thus, it seems likely that neurosecretory axons continually grow and that microglia recognize terminals that obtrude through the perineuronal basal lamina and remove them to maintain a clear perivascular space. This axonal growth may increase at the same time as the cells hypertrophy during increased magnocellular neurosecretory activity. What signal causes the microglia to recognize the protruding terminals and to phagocytose them remains to be discovered. It will be interesting to discover whether the perivascular macrophages show similar changes in chronic salt loading, pregnancy/parturition and lactation. This study was supported by an MRC studentship and Oxford University. We thank D. Hardiman for expert animal care, A. Loreto-Gardner for typing the manuscript, and L. Scott and S. Rodgers for technical assistance. Professor S. Gordon, and Drs. V.H. Perry and L.J. Lawson provided antibodies and helpful advice. Ill Armstrong~ W.E., Tian, M. and Reger, J.F., Elevated extracellular potassium is associated with a reduced extracellular space in rat neural lobe in vitro, J. Neurocytol., 20 (1991) 564 572. [2] Beagley, G.H. and Hatton, G.I., Rapid morphological changes in supraoptic nucleus and posterior pituitary induced by a single hypertonic saline injection, Brain Res. Bull., 28 (1992) 613 618. [3] Blanco, E., Pilgrim, C., Vazquez, R. and Jirikowski, G.F., Plasticity of the interface between oxytocm neurons and the vasculature in late pregnancy: an ultrastructural morphometric study, Acta Histochem.,91 (1991) 165 172. [4] Chapman, D.B., Theodosis, D.T., Montagnese, C., Poulain. D.A. and Morris, J.F., Osmotic stimulation causes structural plasticity of neurone glia relationships of the oxytocin but not vasopressin secreting neurones in the hypothalamic supraoptic nucleus, Neuroscience. 17 (1986) 679 686. [5] Dellmann, H.-D., Stoeckel, M.E., Porte, A. and Stutinsky, F., Ultrastructure of the neurohypophysial glial cells following stalk transection in the rat, Experientia, 30 (1974) 1220 1222.
[6] Duchen, L.W., Changes in the volume of the lobes of the pituitary gland and in the weight and water content of organs of rats given hypertonic saline, J. Endocrinol., 41 (1968) 593-600. [7] Hatton, G.I., Pituicytes, glia and control of terminal secretion, J. Exp. Biol., 139 (1988) 67-79. [8] Hatton, G.I., Bicknell, R.J., Hoyland, J., Bunting, R. and Mason, W.T., Arginine vasopressin mobilises intracellular calcium via V I receptor activation in astrocytes (pituicytes) cultured from adult rat neural lobes, Brain Res., 588 (1992) 75 83. [9] Lawson, L.J., Perry, V.H. and Gordon, S., Microglial responses to physiological change: elevation of DNA synthesis without significant activation in the neurohypophysis of the osmotically stressed mouse, Neuroscience, 56 (1993) 929 938. [10] Leng, G. and Shibuki, K., Extracellular potassium changes in the rat neurohypophysis during activation of the magnocellular neurosecretory system, J. Physiol., 392 (1987) 97-112. [11] Luckman, S.M. and Bicknell, R.J., Morphological plasticity that occurs in the neurohypophysis following activation of the magnocellular neurosecretory system can be mimicked in vitro by fl-adrenergic stimulation, Neuroscience, 39 (1990) 701 709. [12] McLean, I.W. and Nakane, RK., Periodate-lysine-paraformaldehyde fixative: a new fixative for immuno-eleetron microscopy, ,1. Histochem. Cytochem., 22 (1974) 1077-I083. [13] Olivieri-Sangiacomo, C., On the fine structure of the perivascular cells in the neural lobe of rats, Z. Zellforsch., 132 (1972) 25 34. [14] P6rez-Figares, J.M., Fernandez-Llebrez, R, Pdrez, J. and MarinGir6n, F., Presence of microglia-like cells in the neurohypophysis of normal and dehydrated rats, Acta Anat., 127 (1986) 195 200. [15] Perlmutter, L.S., Hatton, G.I. and Tweedle, C.D., Plasticity in the in vitro neurohypophysis: effects of osmotic changes on pituicytes, Neuroscience, 12 (1984) 503 511. [16] Pow, D.V., Perry, V.H., Morris, J.F. and Gordon, S., Microglia in the neurohypophysis associate with and endocytose terminal portions of neurosecretory neurons, Neuroscience, 33 (1989) 567 578. [17] Robinson, A.R, White, T.M. and Mason, D.W., Macrophage heterogeneity in the rat as delineated by two monoclonal antibodies MRC OX-41 and MRC OX-42, the latter recognizing complement receptor type 3, Immunology, 57 (1986) 239 247. [18] Starkey, RM., Turley, L. and Gordon, S., The mouse macrophage-specific glycoprotein defined by monoclonal antibody F4/ 80: characterization, biosynthesis and demonstration of a rat analogue, Immunology, 60 (1987) 117122. [19] Tweedle, C.D. and Hatton, G.I., Glial cell enclosure of neurosecretory endings in the neurohypophysis of the rat, Brain Res., 192 (1980) 555 559. [20] Tweedle, C D . and Hatton, G.I., Evidence for dynamic interactions between pituicytes and neurosecretory axons in the rat, Neuroscience, 5 (1980) 66t 667. [21] Tweedle, C.D. and Hatton, G.I., Magnocellular neuropeptidergic terminals in neurohypophysis: rapid glial release of enclosed axons during parturition, Brain Res. Bull., 8 (1982) 205 209. [22] Tweedle, C.D. and Hatton, G.I., Synapse formation and disappearance in adult rat supraoptic nucleus during different hydration states, Brain Res., 309 (1984) 373 376. [23] Tweedle, C.D. and Hatton, G.I., Morphological adaptability at neurosecretory axonal endings on the neurovascular contact zone of the rat neurohypophysis, Neuroscience, 20 (1987) 241 246.