masts W“ \l unmm l Michfiemm 53%!“ l Waivers?” ‘ —.—_._..,- - This is to eerlil'y that the lhcsheenlhled LACTATION-ASSOCIATED REDISTRIBUTION OF THE GLIAL FIBRILLARY ACIDIC PROTEIN WITHIN THE SUPRAOPTIC NUCLEUS OF THE RAT: AN IMMUNOCYTOCHEMICAL STUDY presented by Adrienne Kay Salm has been accepted towards fulfillment of the requirements for .____.M: A - degree in Psychology. - gab/m Major professor L 0-7639 MSU is an Affirmative Action/Equal Opportunity Institution MSU RETURNING MATERIALS: Place in book drop to LJBRARJES remove this checkout from “ your record. FINES will be charged if book is returned after the date stamped below. ‘r w l’ “-5 Q. _. 3‘1: ;; i ufi‘ .. ‘n—‘ n -—"—h- LACTATION-ASSOCIATED REDISTRIBUTION OF THE GLIAL FIBRILLARY ACIDIC PROTEIN WITHIN THE SUPRAOPTIC NUCLEUS OF THE RAT:AN IMMUNOCYTOCHEMICAL STUDY By Adrienne Kay Salm A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1983 ABSTRACT LACTATION-ASSOCIATED REDISTRIBUTION OF THE GLIAL FIBRILLARY ACIDIC PROTEIN WITHIN THE SUPRAOPTIC NUCLEUS OF THE RAT:AN IMMUNOCYTOCHEMICAL STUDY By Adrienne Kay Salm Previous electron microscopical evidence has demonstrated a lactation or dehydration-induced disappearance of glial processes normally interposed between magnocellular neuroendocrine cell (MNC) somata. These data suggested the hypothesis that glia actively retract their processes when MNCs are highly active. This hypothesis was investigated at the light microscopic level in the supraOptic nuclei (SON) of lactating and estrous rats. The distribution of components of the glial cytoskeleton was visualized using antibodies against the glial fibrillary acidic protein (GFAP). The lateral hypothalamic area (LHA) was also sampled. Computerized image analysis was employed to assess changes in the distribution of this protein concomitant with lactation. Statistical analysis revealed a redistribution of GFAP in the SON of lactating, as compared to estrous, animals which reflected a reduction in the number of glial processes. No such changes were found in the LHA. These results demonstrate a morphological lability of astrocytes in response to a physiological stimulus. This is dedicated to the late Dr. David Karl Bliss, who first encouraged me to do research. Acknowledgments Many thanks, first, to Dr. Glenn I. Hatton for his support, guidance, encouragement and faith, during the development of this thesis. Also, recognition and thanks are due to Mr. Kenneth Smithson for general technical assistance throughtout the project. In addition, his astute analysis of the problems inherent in computerized image analysis, and his ability to surmount them by creating excellent software, made the use of such technology possible for this thesis. Acknowledgment is also due Dr. Gajanan Nilaver for the generous gift of his anti-GFAP serum. I am grateful also to Dr. Richard Dubes for the use of the facilities at the Pattern Recognition and Image Analysis Laboratory at MSU. Dr. William Bukowski deserves a special thanks for contributing his time and expertise in executing the BMDP statistical program used for analysing much of the data. I'm also grateful to Dr. John Gill for illuminating the utility of split-plot designs in biological research. I'd like to thank my other committee members, Dr. Antonio Nunez, Dr. Charles Tweedle, and Dr. Charles Wilson for comments, critiques, and moral support. Also, Dr. Peter Cobbett has been immeasurably helpful with editorial and scientific comments on an earlier draft of the manuscript. Janice Harper, Betty Simon, and Carla Tiller have my undying gratitude for their expert secretarial assistance. Finally, to my family, particularly my parents, and friends, thank-you for having faith and supporting me in all of my efforts. This research was supported by NIH grant # 09140. iv TABLE OF CONTENTS List of Tables..........................................vi List of Figures.......................................viii List of Abbreviations....................................x Introduction.............................................1 Methods..................................................6 Results.................................................20 Discussion..............................................41 List of References......................................47 Appendix A (Protocols)..................................56 Appendix B (Supplies and Equipment).....................66 Appendix C (Data tables, Statistical formulas and Raw data)00000000......0....0....0.0.0.000000000000071 LIST OF TABLES Table 1. Preliminary nucleus circularis data............14 Table 2. Mean Staining Densities-SON....................26 Key to ANOVA and Cell Mean tables.......................71 3a. 3b. 30. 3d. 4a. 4b. 4c. 4d. 4e. 4f. 5a. 5b. 6a. ANOVA Table—Sample Size—Coronal SON.................72 ANOVA Table-Sample Size-Coronal LHA.................73 ANOVA Table-Sample Size-Horizontal SON..............74 ANOVA Table-Sample Size-Horizontal LHA..............75 ANOVA Table-Mean Density-Coronal SON................76 ANOVA Table-Mean Density-Coronal LHA................77 ANOVA Table-Mean Density-Horizontal SON.............78 ANOVA Table- Mean Density-Horizontal LHA............78 Cell Means and Standard Deviations -Mean Density—Coronal LHA...........................79 Cell Means and Standard Deviations -Mean Density- Horizontal SON.......................80 ANOVA Table-Standard Deviation —Coronal SON........................................81 Cell Means and Standard Deviations- Standard Deviation-Coronal SON......................82 ANOVA Table-SkewneSS-HOPiZOl’ltal SON. o o o o o o o o o o o o o o o 083 LIST OF TABLES-cont. 6b. Cell Means and Standard Deviations- Skewness-Horizontal SON.............................84 6c. ANOVA Table-Skewness-Horizontal LHA.................85 6d. Cell Means and Standard Deviations- Skewness- Horizontal LHA............................86 7a. ANOVA Table-Kurtosis-Coronal SON....................87 7b. Cell Means and Standard Deviations- Kurtosis-Coronal SON........................ ..... ...88 Statistical formulas for skewness and kurtosis..........89 Raw Data-OOOOOOOOOOOOOOO0.000..OOOO...0.0.0.000000000000090 LIST OF FIGURES FIGURE PAGE 1. Photomicrograph of a 15 um section through the nucleus circularis. The section has been immunocytochemically stained using antibodies against GFAP. Large arrows point to glial processes. Small arrow points to neuronal cell bodies counterstained with thionine............12 Comparing the frequency distribution characteristics indicates the extent and direction of changes in staining density. A. Astrocyte prior to the addition of a process-inducing substance: little GFAP staining; B. After addition of a process- inducing substance process formation begins: GFAP staining increases; C. Process formation complete: intense GFAP staining...........21 Graph illustrating the mean staining densities of the SON and LHA in coronal control and lactating subjects. The mean staining density was found to be significantly less in the SON of coronal lactating animals as compared to controls. No changes were found to occur in the LHA under the same conditions.*=p<.05...........25 . Frequency histogram of GFAP staining densities from subject #91; a ten day lactating animal. A micrograph of the anterior SON of this subject is displayed in figure 5...............................29 Coronal section from the anterior SON of subject #91, a ten day lactating animal. Note that the paucity of staining is confined to the nucleus. Arrows delineate the SON. Higher power micrograph of section shown in figure 8................................... ...... 29 Frequency histogram of GFAP staining densities from subject #13; a control animal.. A micrograph of the anterior SON of this subject is displayed in figure 7........................................31 \dii LIST OF FIGURES-—Continued FIGURE PAGE 7. Coronal section from the anterior SON of subject #13, a control animal. Arrows delineate the SON. Higher power micrograph of section shown in figure 9........................31 8. Lower power micrograph of a section immunostained for GFAP from a lactating animal. Note the paucity of staining in the SON and compare to the SON shown in figure 9................33 9. Lower power micrograph of a section immunostained for GFAP taken from a control animal. Notice the dense staining in SON and compare with that in the SON shown in figure 8.................................... ..... 33 10. Densely stained ventral pial-glial plexus, SON. Horizontal section, control animal.............37 11. Dorsal SON from same subject as figure 10. Horizontal sectionOOOOOOO...OOOOOOOOOOOOOOOOOOOOO00.37 12. Densely stained ventral pial-glial plexus, SON. Horizontal section, lactating subject... ..... ..39 13. Dorsal SON from same subject as figure 12. Horizontal sectionOOIOOOO...O00....0.0.00.000000000039 LIST OF ABBREVIATIONS 3V THIRD VENTRICLE BV BLOOD VESSEL GFAP GLIAL FIBRILLARY ACIDIC PROTEIN LHA LATERAL HYPOTHALAMIC AREA NC NUCLEUS CIRCULARIS OT OPTIC TRACT PG PIAL-GLIAL PLEXUS SCN SUPRACHIASMATIC NUCLEUS SON SUPRAOPTIC NUCLEUS INTRODUCTION In recent years the list of functions attributed to neuroglia has grown to include much more than merely being, as Virchow (1840) named them, the "glue" of the nervous system. It has been repeatedly demonstrated both lE.XlE£2 and in viva that astrocytes are capable of modifying and responding to their enviroments (see Varon and Somjen, 1979, for review). Ultrastructural, morphological, and biochemical changes have been variously observed in astrocytes to be associated with mechanically produced CNS injuries (Bignami and Dahl, 1973; Latov et al.,1979)3 irradiation (Maxwell and Kruger,1965); topical alumina (Harris, 1973); nerve sectioning (Vaughn and Pease,1970); hypertension (Hanakita et al.,1980); and hypoxia (Landis and Reese, 1981). Laursen and Diemer (1980) have shown that ultrastructural changes also occur in astrocytes during increased CNS ammonium levels created by portocaval anastomosis. At the light micros00pic level, astrocytic plasticity has been demonstrated in response to pharmacological manipulations. Morphological changes have been induced in immmature astrocytes in primary monolayer cultures by the addition into the medium of dibutryl cyclic adenosine monophosphate (Kimelberg et al., 1980; Manthorpe et al., 1979; Moonen et al, 1976); norepinephrine (Narumi et al., 1 2 1978; Steig et al., 1980) and prostaglandins (Tardy et al., 1981). Such changes are positively correlated with the polymerization into filaments of the glial fibrillary acidic protein (GFAP) which exists in these cells. GFAP is a 10 nm intermediate filament protein with a molecular weight of 49-55 daltons which is a major component of the mature astrocytic cytoskeleton (Liem, 1982). This protein has received much attention in recent years because it serves as a "marker" of cells which are astroglial in origin (Bignami et al., 1980) and because its production and/or subsequent aggregation into filaments is often concomitant with other astrocytic changes associated with neurOpathologies. In this regard, GFAP has become relevant to clinicians because its presence is associated with cerebral carcinomas (Duffy et al., 1978;1979; DeArmond et al., 1980 for review) fetal abnormalities (Pertti et al. 1980), and Alzheimer-type senile dementia, where both an increase in the amount of immunostainable GFAP (Duffy et al.,1980) and in the number of GFAP immunopositive astrocytes in cerebral cortex (Schechter et al., 1981) have been described. Possible role 2: glia in neurosecretion. Despite the growing evidence that neuroglia and GFAP are able to respond to a variety of stimuli, the current literature is, with a few exceptions, notably lacking in examples of glial lability in gitu when conditions are both physiological and non-pathological. These exceptions include a number of descriptions which indicate that glial cells 3 within the hypothalamo-neurohyp0physial neurosecretory system of the rat undergo biochemical changes which appear concurrently with the release of neurohypophysial hormones. Watson (1971) found that perfusion of the cerebral ventricles with a solution containing a high concentration of either potassium or phOSphate, or the calcium chelator, EDTA, caused a marked and reversible increase in dry mass of hypothalamic paraventricular nucleus (PVN) astrocytes and nearby ependyma. He suggested that this was due to increased RNA and protein synthesis in these cells. In a similar study using interference microscopy and ultra-violet absorbtion microspectrography, Watson (1972) found that the dry mass of supraoptic nucleus (SON) and PVN astrocytes transiently increased with lactation or one day of dehydration. In the SON and neural lobe, 14 days of substituting a 1.75% NaCl solution for drinking water, resulted in dramatic increase in the number of astrocytes labelled by injections of 3-H-thymidine (Paterson and LeBlond, 1977). These authors reported, however that the over-all number of astrocytes/ 100 neurons remained constant suggesting that some cell death occurred. Further evidence, although indirect, for a role of astrocytes in neurosecretion came from LaFarga et al. (1975) who were perhaps the first investigators to specifically describe the lack of a glial barrier between dehydration activated magnocellular neurons. Using electron microscopy (EM), these investigators also described unusually long membrane appositions between neurosecretory somata and further suggested that junctions similar to gap 4 junctions might exist between these as a basis for synchronous electrical activity. More recently, at the EM level, Tweedle and Hatton (1976) also found that the conditon of dehydration was accompanied by a disappearance of glial elements which are normally interposed between the somata of neurons in the nucleus circularis (NC) and the SON. An active retraction of glial processes was hypothesized to account for the significant increases in soma-somatic apposition as these changes occur prior to the increased cell volume observed after brief dehydration (Hatton and Walters,1973; Hatton,1976). The reappearance of glial processes between neurosecretory somata with subsequent rehydration, provides further support for the participation of astrocytes in osmoregulation (Tweedle and Hatton, 1977). The supposed retraction of fine glial processes has also been shown to occur within the SON and NC in lactating rats ( Hatton and Tweedle, 1980; 1982). When compared to virgin females, Hatton and Tweedle found that the percentage of cells showing cell-cell contacts increased twelve-fold, from 3.5% to 44%. The increase in total membrane apposition was an even more dramatic 41-fold over that of the virgin females. A similar pattern of change was observed in NC, however to a lesser extent. Theodosis et al.(1981) also reported an apparent glial retraction from between the SON somata in lactating animals. Using a somewhat different sampling method, they calculated a five-fold increase in total surface membrane contact between SON cells of 5 lactating rats. The data from these latter two investigations indicate that lactation is a more effective stimulus than dehydration in promoting the disappearance of glial processes from between magnocellular neurons. In fact, the additional hormonal demand of one day of dehydration actually reduced the magnitude of this phenomena (Theodosis et al., 1981). Elsewhere in this neurosecretory system, in the neurohypophysis (Tweedle and Hatton, 1980 a & b;1982) and in the PVN (Gregory, Tweedle and Hatton, 1980) glial retraction has likewise been described in association with the lactation and dehydration stimuli. Statement 2: problem. Thus far the evidence for changes in glial morphology has been indirect i.e., the absence of glial processes where they are normally observed. However, detecting changes in glial morphology is difficult at the high magnifications used in EM. Preliminary light level investigations using peroxidase anti-peroxidase (PAP) immunocytochemistry (ICC) for GFAP have shown an intricate glial reticulum within the SCN (Salm and Hatton,1980) and NC in normally hydrated animals. The dense packing of neurons in these nuclei permits the suggestion that a change of the glial elements, be it a shortening of the processes or an aggregation of cyt0plasmic GFAP into filaments could conceivably result in an increased neurosecretory cell apposition. In addition, astrocytes of both nuclei are invariably associated with vasculature and those of the SON occupy a particularly close 6 Proximity to the arachnoid space, although whether any contact with CSF occurs is not known. Therefore, the glia of these nuclei appear to be positioned so as to be sensitive to osmotic or other cues in the blood and/or CSF, and to perhaps influence neurosecretory responses via changes in their own morphology. It may be possible to directly demonstrate that morphological changes of astrocytes have a role in the day to day running of the nervous system by documenting changes in these cells themselves. To this end, a method was develOped to determine and compare the pattern and distribution of their cytoskelatal protein, GFAP, as visualized by PAP ICC with primary antibodies directed against GFAP. This was done in virgin estrous, and lactating rats. METHODS Subjects. Subjects were twelve female Holtzman rats, 94- 98 days of age, divided into two groups of six: virgins in the estrous phase of their reproductive cycle ( hereafter referred to as controls), and lactating animals 8 or 10 days post partum. All of the nursing animals had full litters, each nipple potentially being occupied during suckling. Animals were housed in a 12:12 light cycle with lab chow and water available ad libitum. Procedure. Rats were killed within four hours after the onset of the dark phase of the light cycle, a time when most of (prolactin excepted) estrous cycling hormones in estrous 7 animals are relatively low (Nequin,1979). Time of estrus was determined by vaginal lavage yielding cornified epithelium. Animals were followed through at least two cycles and those having consistently equivocal smears were not used. It appeared important to control for the phase of the estrous cycle as there is a growing body of evidence which indicates estrogens may act to modify neurosecretory activity. Skowsky, Swan and Smith (1979) have reported an estrogen replacement stimulated rise in serum vasopressin in ovariectomized rats. Combined autoradiography and ICC indicate that estrogen receptors are localized to hormone containing cells of magnocellular nuclei (Sar and Stumpf, 1980; 1981; Rhodes, Morrell and Pfaff, 1981). Behavioral data indicate that volemic thirst appears to be influenced by the estrous cycle, being least apparent during estrus (Findlay, Fitzsimmons, and Kucharczyk, 1979). Negoro et al. (1973) and Freund-Mercier and Richard (1977) have found that PVN neurons that project to the posterior pituitary vary in their spontaneous and stimulation induced firing rates over the course of the estrous cycle. By monitoring neurosecretory activity during normal reproductive cycles in ovariectomized animals with and without estrogen replacement and during pregnancy and lactation, Swaab and Jongkind (1970) found that the greatest level of activity coincided with (reported) high levels of plasma gonadotrOphs. However, no clear picture has yet emerged as to the exact relation between the gonadotrophins and neurohypOphysial secretory activity. Histology. Following transcardiac perfusion with 1% glutaraldehyde, 2.5% paraformaldehyde in 0.1M cacodylate buffer (pH 7.25), brains were dissected from the skull, trimmed into hypothalmic blocks, immersed in fixative until block was firm, and rinsed extensively in Tris buffered saline (see Appendix A) for 48 hours. At this time, a randomly selected code number was assigned to each brain by an associate, in order that subsequent operations be performed without knowledge of the experimental group membership. The block was dehydrated in ethanol and embedded in polyethylene glycol for 15um sectioning with a rotary microtome, according to the procedure of Smithson et al.(1983). Sections were cut in either coronal or horizontal planes with three brains per orientation per group. Every third section from the coronal group, and all sections from the horizontal group were collected and rehydrated for immunocytochemical processing with a primary antiserum raised against GFAP. Additional sections were processed from the coronally cut brains through areas identified as containing NC. Immunocytochemistry For immunocytochemical processing, sections were rehydrated in individual wells of a Plexiglas template which requires 100ul of a given solution to adequately cover the tissue. In order to eliminate non-specific staining and allow for clearer visualization of stained processes a pretreatment (Schachner et al.,1977) was first applied to 9 each section. Subsequently, the peroxidase-antiperoxidase immunostaining procedure of Sternberger (1974) was employed using primary antibodies against GFAP purified from human cerebral white matter (Latov et al., 1979). In order to control for the inevitable variability which occurs between "batches" of immunocytochemically stained tissue, control and experimental sections were reacted simultaneously, with the exception of sections from one control and experimental brain each. In addition, the timing of each step of the protocol was held constant across "batches" of tissue. It was thus attempted to minimize method-induced variabilty. Details of the pretreament and ICC protocols are given in Appendix A. After ICC was completed, sections were mounted, air dried, dehydrated in alcohol and couterstained with a 0.5% thionine solution. Following coverslipping with #1.5 coverslips the slides were examined under low magnification with a Zeiss microsc0pe. For analysis of the SCN of coronally sectioned brains 10 sections from the anterior and posterior SON each were selected. The boundary between these regions was semi-arbitrarily determined to be between the caudal part of suprachiasmatic nucleus and the most rostal portion of the medial paraventricular subnucleus. Equivocal sections were excluded from analysis. Also, based upon the overall intensity of the thionine stain, sections which were obviously thicker or thinner were omitted from these groups. For analysis of sections in the horizontal plane the five most dorsal and five most ventral sections were chosen. 10 Antiserum production and specificity Antiserum against GFAP and preimmune serum were generous gifts of Drs. Gajanan Nilaver and Earl Zimmerman who prepared the antiserum according to the protocol of Dahl and Bignami (1975). Their procedures are summarized in Appendix A. Preliminary data collection. Preliminary data collection was performed on the NC of two experimental and two control animals. This was done in order to assess the sensitivity of the data collection methodolgy to presumptive changes in the distribution of astrocytic elements between control and experimental animals. NC was chosen for this analysis because of its apparent orderly neuronal arrangement and uncomplicated distribution of glial processes (Figure 1). Methods. All slides containing sections with NC were given individual code numbers by a colleague in order that all measurements be made without knowledge of experimental group. A microscope eyepiece reticule was used in making the following counts at 312.5X magnification: 1.The number of GFAP-positive glial processes appearing to be interposed directly between magnocellular somata. 2. The number of GFAP-positive glial processes visible which are not interposed between somata. These measurments were done in order to ascertain LL Figure 1. Photomicrograph of a 15 um section through the nucleus circularis. The section has been immunocytochemically stained using antibodies against GFAP. Large arrows point to glial processes. Small arrows points to neuronal cell bodies counterstained with thionine. 13 whether the number of visible processes within the nucleus changed with experimental manipulation, and to determine where such changes might be localized in relation to the neurons therein. To quantify these variables, the nucleus was centered within the field covered by the reticule as observed through the eyepiece. The incidence of each variable was counted for each grid square overlying the nucleus or a portion of it. All numerical data were then converted to a value normalized for area, expressed as /cubic um, assuming sections to be ramdomly variable around a mean 15um thickness. Analysis. Preliminary testing of the data caused the hypothesis of equal variance to be rejected but one of equal coefficient of variation to be accepted. Accordingly, (Gill, 1978), the raw data were first transformed to logs prior to performing a standard t-test. This test was applied to detect differences between groups in the number of glial processes interposed between neuronal somata and in the total number of glial processes observed within the nucleus. Examination of the means for these two measures show that the means were virtually identical. 14 Table 1. Preliminary nucleus circularis data Mean processes pot Mean processes between interposed /um NC somata /umFNC CONTROL -3.62 i .63 -4.03 :_.22 LACTATING -3.76 + .30 -3.98 + .41 Subsequent t-tests on these two measures confirmed that there were no significant differences between control and lactating animals. Since changes in the amount of neurosecretory cell apposition had been shown to occur, both during dehydration and lactation (Tweedle and Hatton, 1977; Hatton and Tweedle, 1982), it was surprising that this was not reflected by changes in the number of observable processes in this nucleus. This finding prompted an evaluation of the methodology employed. Possible sources of the method's failure to detect differences were considered to be the following: (a) no changes actually occur under these conditions, and (b) the actual number of astrocytic processes don't change, but their dimensions do. Possibility (a) was rejected on the basis of accumulated evidence cited in the initial pages of this thesis. Possibility (b) seemed the more reasonable in light of several considerations. First it was obvious that glial processes, as visualzed by PAP ICC for GFAP, vary in thickness (Figure 1). Simply counting the presence or absence of such processes ignores changes which may be 15 occurring in the dimensions of such processes. In addition, it is likely that GFAP ICC does not darkly stain the entire astrocyte. Intracellular injections of horseradish peroxidase (Picker and Goldring,1982) and metallic stains (personal observation) reveal quite a different picture of these cells. These techniques invariably reveal extensive velate material between the main processes. Such velate material is not invisible in immunostained astrocytes, but is seen as gradations of a light brown "wash" which is difficult to differentiate from any background staining and is still more difficult to quantify. It was for these reasons that a more sensitive method was sought. Such a method would have to exceed the capabilties of the previously employed methodology in these respects: improved resolution, i.e., able to discriminate minute differences in process width; improved ability to discriminate fine gradations in immunostaining densities and an ability to objectively and consistently quantify these differences. Such precision is found in computerized image analysis systems. One such system, the Eyecom II, is available at Michigan State University and was therefore used for this project. The images were produced by photographing each section through a microsc0pe using a black and white negative film. The negatives were then scanned by the Eyecom II for image analysis. Producing the image Optics. Before beginning photography, every precaution must be taken to Optimize resolution with the available 16 equipment. Details concerning equipment and procedures are in Appendix A. Photography. The major considerations here were the choices of film and developer, with special reference to the range of contrasts or densities which the film is capable of rendering with a certain developer. Since the Eyecom II system has a dynamic range of 2.2 density units, a fine grain (high resolution) film which could be developed to a similar density range was sought. Kodak Technical Pan 2415 35 mm negative film was chosen due to its high resolution (320 lines/mm) and compatible density range of 1.75 density units when develOped with Kodak HC-110, dilution D. Also, this panchromatic film has the attractive feature of extended red sensitivity which makes it especially suited for depicting the reddish brown reaction product of DAB. To further enhance contrast, a #80 blue filter was used. Another important factor related to the actual photography was holding constant the illumination of the specimen. This task required that not only the amount of light entering the system be the same, but also that it pass through each section in the same way, i.e., the field diaphragm and the substage condenser be similarly adjusted. Illumination was held constant by metering the light transmitted through the Permount adjacent to each section using a Nikon light meter. The light sensor was held to a Nikon focusing eyepiece connected in series to the mechanical tube and camera. The voltage of the light source was then adjusted to give a 6uA i .5 uA reading on the light 17 meter. Prior to photography, the substage aperature was adjusted to 80% of its maximum as recommended for modified Kohler illumination. This position was marked and checked throughout the photography process. Random numbers from 1-220 were generated without replacement and assigned in order of generation to each section/animal to be sampled. These numbers were then arranged in order from 1-220 for photography. Thus, the order in which sections were photographed was randomized. A summary of the procedures employed for photography appears in Appendix A. Data collection. Data were obtained from each negative by placing it on the backlighted table directly underneath the Eyecom II Scanner. In order to keep unwanted table light from entering the Eyecom II, a posterboard template was constructed which entirely covered the light table except for a 2" by 4" portion in the center where the slide was placed. The image from the negative was then transmitted to the video display terminal where a joystick cursor was used to encircle the area to be sampled. A gray scale frequency distribution was subsequently determined for each sample. Data from negatives from a single roll of film were collected as a group. Also included on each roll were two frames of neutral density filters which represented two points of the density spectrum (63% and 1% transmittance). These were used to calibrate the camera in order that the gray scale distribution be placed in the same relative 18 position for each role of film. This compensated for possible differences in film "batches" and film development. On each negative a central portion of the SCN was sampled which was interpreted by the Eyecom II system as a frequency distribution of staining densities. For control purposes, a portion of the lateral hypothalamic area was also sampled. Since technical considerations necessitated removal of the thionine stain, NC was not sampled for this analysis, as it was not visible without counterstain. From the density histograms the distributional features of mean, standard deviation, skewness, and kurtosis, were determined. Data from sections representing each location within SON were combined to give a composite distribution representing that location for each animal. The distributional features mentioned above were then statistically analysed as the dependent variables. Statistical analysis Data were analysed separately for horizontally and coronally sectioned brains using a split-plot design for repeated measures over space (Gill,1978,vol.2).The model for this design is: 19 Statistical Model (Split-Plot design) Yijk = u + oi + D(i)j + 8k + (as)ik + (D8)(i)jk + E(ijk) where ai = treatment effects (lactation vs. estrus) D(i)j = random effects of sUbjects Bk = effects of location (a8)ik = interactions of treatments with locations (D8)(i)jk = interactions of subjects with location (Not separable from experimental error). Analysis of these variables was accomplished with the BMDP statistical package PV2. This program tests the ANOVA assumptions and fills in data for missing cases by calculation of harmonic means. This was necessary for one horizontal control animal as this tissue was lost in processing. RESULTS Theoretical considerations. When morphological changes are induced in astrocytes ‘in 11:59 by the additon into the medium of dbcAMP, process formation is observed (Kimelburg et al.,1980). If immunostaining for GFAP is done at intervals during process formation, an increasingly higher density staining is produced, which corresponds to filament formation (Trimmer et al., 1979). If a frequency histogram of densities is obtained at the same intervals during the transformation, certain distributional features would be expected to change. Figure 2 illustrates the postulated relationship between the shape of the frequency distribution and the state of the GFAP immunostain. Methodological controls. Sample size. The data for sample size are listed in Tables 3a-d, Appendix C. All comparisons of the number of picture elements (pixels) sampled revealed that there were no significant differences in sample size (F<1) for both treatments and locations. This was true in both the horizontal and coronal planes. These data indicate that a relatively consistent sampling of the SON was achieved. Mean staining density—LHA. Data for the mean staining density of the LHA in coronal section are presented in Table 4b, Appendix C. Analysis of this variable in coronal 20 21 Figure 2. Comparing the frequency distribution characteristics indicates the extent and direction of changes in staining density. A. Astrocyte prior to the addition of a process-inducing substance: little GFAP staining; B. After addition of a process-inducing substance process formation begins: GFAP staining increases; C. Process formation complete: intense GFAP staining. 22 w. 1:3 0% ISISITV "0% IIISHY "0‘ D73: fl IQIEICV 0% uum 1.0% Figure 2. 23 sections revealed virtually identical mean densities of control and lactating rats in the LHA (Figure 3). Similar results were obtained in horizontal sections (Table 4d, Appendix C). These data indicate successful production of a comparable immunostain in the LHAs of both lactating and control animals. Tests 9f treatments, locations, and interactions 2: treatments with locations. In order to facilitate the reader's understanding of these results, some interpretations have been included. Differences in mean, standard deviation, skewness and kurtosis have been related to the pattern of the GFAP immunoreactivity which is quantified by these measures. Mean staining densities: SON. Table 2 lists the data for mean density level of the SON for control and lactating subjects in coronal section (ANOVA Table 4a, Appendix C). Slit-plot analysis of variance revealed that there were significant differences between control and lactating groups (p<.05, F=8.12, dfs=1,3) on this measure in coronally sectioned tissue. SONs from lactating rats showed a smaller mean staining density. These data indicate a per unit area reduction in the immunostainable glial cytoskeletal protein. As with all immunological methods, such changes can be related to the availabilty of antigenic sites for binding with the primary antibody. The changes in staining distribution observed here therefore reflect a reduction in available binding sites brought about by a 24 Figure 3. Graph illustrating the mean staining densities of the SON and LHA in control and lactating subjects. The mean staining density was found to be significantly less in the SONs of coronal lactating animals as compared to controls. No changes were found to occur in the LHA under the same conditions. *=p<.05. 25 '0 (fl - control , llllllllllllll um CD N a) 00 MEANDENSIW m a) LLLILLILLILILILIHLLLIILJ as _,. '1' so _1 77 SON LHA Figure 3. 26 No.mm H©.mw «0.0m om.mw com: Cowumooq mw.ww mm.om 0H.mw ma.~w mm.Nw um.Hw wdwumuomd do.ma w u nonnm Unmpcmum 20m I mmfluwmcoa wcficwmum coo: on.mm com: udmfiummuH om.¢m amwuam> mA02m30mm... 1500 258 123 PIXEL GRAY VALUE STAINING DENSITY Figne 4. Fugue 5. Figure 6. Frequency histogram of GFAP staining densities from subject #13; a control animal. A micrograph of the anterior SON of this subject is displayed in figure 7. Figure 7. Coronal section from the anterior SON of subject #13, a control animal. Arrows delineate the SON. Higher I. power micrograph of section shown in figure 9. 31 7... CONTROL FREQUENCY a 8 l 1500 .1 l l l 1 12.8 255 PIXEL GRAY VALUE STAINING DENSITY —: Fignz 7. 32 Figure 8. Lower power micrograph of a section immunostained for GFAP from a lactating animal. Note the paucity Of staining in the SON and compare to the SON shown in figure 9. Figure 9. Lower power micrograph Of a section immunostained for GFAP taken from a control animal. Notice the dense staining in SON and compare with that in the SON shown in figure 8. 33 Figure 8 . Figure 9 . 34 significant (p,<01, F=14.25, dfs=1,3) difference was found for the location effects within coronally sectioned SONs on this measure. Examination Of the mean standard deviation for each location shows that the anterior portion of the control SON has a much greater variabilty Of staining densities than does the posterior portion. For this measure, "variability" reflects the frequency with which extreme density values occur. This suggests that the glial protein in the anterior SON exists in a wider range of configurations than it does in posterior SON. Interestingly, a strong trend towards an interaction Of treatments with locations was evident in the predominantly oxytocinergic anterior SON (p=.07,F=5.61, dfs=1,3). In control tissue the anterior SON showed a greater variabilty than did posterior SON (Means: 29.9 and 24.1 respectively). However, anterior SON variabilty decreased markedly in lactating animals whereas it stayed constant in the posterior SON. These data can be interpreted as evidence that lactation is accompanied by a morphological homogeneity in the glia of anterior SON. Also, this decrease in variability supports the interpretation that the decreased staining density seen in lactating animals is due to an overall decrease in density, rather than the formation of a few high density areas; as these latter changes would increase the frequency Of a wide range Of staining densities, i.e., increase variability. 35 The third momentzskewness. Data are listed in Tables 6a-d, Appendix C for this variable. The measure Of skewness indicates where the majority Of immunostaining density values in the distribution are clustered. While not reaching significance, (p=.06, F=8.12, dfs=1,3; P=.O9, F=5.65, dfs=1,3) trends were once again noted within the horizontal SON for location and interaction effects respectively. The difference in location effects can be easily Observed in Figures 10—13. In control tissue the dorsal SON staining is clustered farther to the left (.735) than are those of the ventral SON (.069). This indicates that more Of the area sampled of dorsal SON was less densely stained than in ventral SON. This is precisely what one would expect in the ventral hypothalamus which contains the darkly staining pial-glial plexus. The trends noted for interactions are interesting in that they again demonstrate a regression towards the homogeneity of staining in the lactating animals which was evident on the standard deviation measure. Staining is redistributed to a skew Of .425 vs. .364 for dorsal and ventral SON respectively. The fourth moment. Data for this measure are displayed in Tables 7a and b, Appendix C. Again, a srong trend (p=.06,F=6.20,dfs=1,3) towards treatment differences for kurtosis, which is a measure Of the peakedness Of the staining distribution, was Observed in coronal SON. Comparisons Of the means shows Figure 10. Densely stained ventral pial-glial plexus, SON. Horizontal section, control animal. Figure 11. Dorsal SON from same subject as figure 10. Horizontal section. 37 Figure 10 . Figure 11 . 38 Figure 12. Densely stained ventral pial-glial plexus, SON. Horizontal section, lactating subject. Figure 13. Dorsal SON from same subject as figure 12. Horizontal section. Figure 13. 40 that the staining distribution is more leptokurtic in the lactating tissue. This is evident in both anterior and posterior SON, but is particularly dramatic in anterior SON, the two portions showing a 1074% and 225% increase over controls respectively. The increased peakedness reflects a decrease in the range Of staining densities i.e. more homogeneity Of staining. Summary Statistical analysis revealed an overall redistribution of GFAP in the supraoptic nucleus of lactating animals. Although Observed in all parts Of the nucleus, the changes were far more dramatic in the predominantly oxytocinergic portions of the SON. In addition, the changes were to a less dense, more homogeneous distribution Of the GFAP, reflective of an absence of high density staining characteristic Of that Observed in glial processes. DISCUSSION The data presented in this report support the view that glia in the SON are capable of altering their morphology during lactation. Although not all of the measures reached conventional levels of statistical significance, the conservative nature of the statistical test which is required by the large number of comparisons made must be considered. In the SON, each significantly and near significantly different measure was congruent with the other to give an overall consistent picture of redistributed GFAP in that nucleus. This was in contrast to the data from the LHA which showed no such redistribution. These results are important for a number of reasons. First, the plasticity Of SON glia has been demonstrated i3 vivo. While GFAP redistribution has been shown in 31:59, it is important to demonstrate that conditions which parallel those Of the culture dish exist in 3112 as well. Second, the stimulus necessary for inducing this phenomenon was a physiological one: presumably parturition and suckling. Ultrastructural lability of glia has been shown in 1119, however it has been induced by CNS insults Of various kinds (Maxwell and Kruger, 1965; Harris, 1973; Vaughn and Pease, 41 42 1970; and Laursen and Diemer, 1980). With the goal of understanding GFAP's role in neuropathologies, 13 11:39 studies have usually been Of cells from pathological tissue, 9. g., astrocytomas (Duffy et al.,1982), senile plaques (Duffy et al.,1980), C6 Gliomas (Mares et al., 1981) and Alzheimer's diseased cortex (Schechter et al., 1981). The demonstration of changes in this protein which occur in normal tissue in response to a natural stimulus implies a normal underlying mechanism which somehow has gone awry in pathological tissue. Third, these data provide direct support for the hypothesis that glia actively retract their processes during conditions Of high hormone demand. While this phenomenon has been inferred many times in EM studies (LaFarga et al., 1975; Hatton and Tweedle 1980; Hatton and Tweedle,1982; Theodosis et al.,1981;and Tweedle and Hatton;1977) this level of analysis has not provided qualitative evidence that the astrocytes themselves actually change. Finally, these data suggest that the "undifferentiated" state which is Often associated with immature astrocytes in culture prior to the addition into the medium Of process-inducing subtances, perhaps may be just one of the morphological forms which an astrocyte may assume during its lifetime. An implication Of this is that the distinction made between fibrous and protoplasmic astrocytes may actually be functional rather than genomic . It is illuminating to review recent work by Duffy and coworkers (1982) who have elegantly correlated astrocytic 43 morphology and plasticity in astrocytic primary cultures. Using time lapse photography and subsequent immunostaining for the GFAP, they have shown that the states Of greatest motility are accompanied by a lack Of glial processes and by the presence of a diffuse, homogeneous, distribution Of the GFAP immunostaining. The appearance Of a similar staining distribution in the SON of lactating animals is consistent with a state Of heightened glial motility in this nucleus during lactation. The consequence (Tweedle and Hatton,1977) of glial retraction is to permit a close 6-8 nm apposition between neurosecretory cells to be formed. It is hypothesized that these close appositions influence the physiology of the neurosecretory cells. First, it may allow the formation Of junctional complexes between these cells. Andrew et al.,(1981) have shown that dye-coupling occurs between magnocellular cells ig‘yitgq in the SON. Recently, it has been shown that the incidence Of dye-coupling between neurosecretory cells in the paraventricular nucleus lg 13339 is influenced by conditions Of high hormone demand (Cobbett and Hatton, submitted). While the relationship Of dye-coupling to the existence of gap junctions is still controversial, the presence Of electrotonic junctions would provide a means Of synchronizing the large number Of oxytocinergic cells which burst simultaneously preceding milk ejection (Wakerley and Lincoln, 1973). A second way that glial retraction may influence neurosecretory cells is 44 by altering the extracellular ionic milieu. It has Often been suggested that astrocytes take up potassium (see Varon and Somjen, 1979 for review) and it is therefore conceivable that when cells are active, the absence Of glia would result in raised extracellular potassium, resulting in an overall increased level Of excitabilty Of the neurosecretory cells. Such interaction of closely apposed neurons has been demonstrated in the cockroach (Yarom and Spira, 1982). Computerized image analysis is a technique which brings precision to the practice of quantifying immunological and histological staining patterns. It Offers improved resolution, consistency, and Objectivity. When information regarding anatomical distribution is desired, (i.e. is the GFAP localized to processes or is it diffused?), this technique Offers a clear advantage over radioimmunoassay. Similar methods have been used (Cowen and Burnstock,1982) to study changes in the pattern and density of catacholamine histofluoresence. While the goal Of this study was purely to document relative differences between groups, Benno and coworkers(1982) have painstakingly succeeded in correlating (r=.96) immunostaining density with a biochemical assay. It is well known that other cytoskeletal proteins are present in astrocytes, and indeed controversy and confusion have existed in the literature concerning the similarities Of GFAP, desmin, vimentin, neurofilament, the microfilament 45 actin, and tubulin (Bigbee and Eng,1982; Bignami et al.,1982; Bray and Gilbert,1981; Ciesielski-Treska et al.,1982,a and b; Dahl et al.,1981,Dahl and Bignami,1982; Geisler et al., 1982; Goldman et al.,1978; Lazarides,1979; Lucas et al.,1980; Reugar et al.,1981; Strocchi et al.,1982; Yen et al.,1976). Suffice it to say, that on the basis Of molecular weights, cleavage products, isoelectric variants, tissue Of origin, and immunostaining patterns, the possible cross—reactivity Of the antiserum used in this study with any Of the above proteins is unlikely under the conditions of this experiment. Redistribution Of both actin and tubulin, however, has been demonstrated in 113:9 concurrent with morphological changes Of astrocytes (Ciesielski-Treska et al.,1982 a & b.) and the participation of these proteins in glial plasticity cannot be ruled out. It should be noted that an intriguing exception to the GFAP being localized to neuroectodermally derived astrocytes has recently been found. Hatfield et al.,(submitted) using both immunological and biochemical methods, have demonstrated that GFAP exists in lens epithelium. These authors put forth the suggestion that GFAP may be induced in this ectodermal tissue by functional demands. The underlying mechanisms which govern the formation and configuration Of the GFAP are unknown. An understanding Of these processes would undoubtedly shed light on the 46 influences Of glia on the release Of neurotransmitters and/or hormones as well as the abberations Of the GFAP which have been Observed in neuropathologies. In culture preparations, several substances have been shown to induce process formation (Narumi et al.,1978; Tardy et al.,1981) although whether they directly influence the production Of GFAP monomer subunits is not clear. Many of these are known to mimic or stimulate adrenergic second messenger systems (Kimelburg et al.,1980; Manthorpe et al.,1979; Moonen et al.,1976; Narumi et al.,1978; Steig et al., 1980). Indeed, astrocyte cultures are routinely used tO study adrenergic receptor mechanisms although functional correlates of these cells have not been demonstrated lg 1119. It's interesting therefore to consider that the SON is richly innervated by catecholaminergic terminals (Katchachurian and Sladek,1980). Whether these terminals directly affect the morphology Of the glia in SON is currently under investigation. LIST OF REFERENCES REFERENCES Andrew, R. D., B. A. MacVicar, F. E. Dudek,and G. I. Hatton(1981) Dye-coupling by gap junctions in magnocellular neuroendocrine cells of the hypothalamus. Science. 211 :1187. BennO,R.H., L. W. Tucker, T. H. JOh, , and D. J. Reis(1982) Quantitative immunocytochemistry Of tyrosine hydroxylase in rat brain 1. Development of a computer assisted method using the peroxidase-antiperoxidase technique. Brain Res. 246 :225-236. Bigbee, J. W., and L. F. Eng(1982) Analysis and comparison Of 13 vitro synthesized glial fibrillary acidic protein with rat CNS intermediate filament proteins. J. Neurochem. 38:130-134. Bignami,A., and D. Dahl(1973) The radial glial of Muller and their response to injury. An immunofluorescence stud with antibodies to the glial fibrillary acidic (GFA protein. Exp. Eye Res. 28 :63-69. Bignami,A., D. Dahl, and D.C. Reuger (1980) Glial fibrillary acidic protein(GFA)iin normal neural cells and in pathological conditions. In Advances in Cellular Neurobiology vol.1, S. Federoff and L. Hertz, eds., pp. 285-310 Academic Press, New York. Bignami,A., T. Raju, and D. Dahl(1982) Localization Of vimentin, the nonspecific intermediate filament protein, in embryonal glia and in early differentiating neurons. l2 vivo and i2 vitro immunofluorescence study Of the rat embryo with vimentin and neurofilament antisera. Developmental Biology, 91 :286-295. Bray, D. and D. Gilbert(1981) Cytoskeletal elements in neurons. In Ann. Rev. Neurosci.,vol.4,W. M. Cowan,Z. W. Hall, and E.R. Kandel,eds.pp.505—523,Annual Reviews Inc.,PalO Alto Ciesielski-Treska,J., B. Guerold, and D. Aunis (1982) Immunofluorscence study on the organization Of actin in astroglial cells in primary cultures. Neuroscience, 1 :509-522. 47 48 Ciesielski-Treska, J., M.-F. Bader, and D. Aunis(1982) Microtubular organization in flat epitheloid and stellate process bearing astrocytes in culture. Neurochemical Research, 1 :275-285. Cobbett, P. J. R., and G. I. Hatton (submitted) Sex and the coupled neuron: dye coupling in hypothalamic slices under basal and hyperosmotic conditions. Cowen, T. and G. Burnstock (1982) Image analysis Of catecholamine fluorescence. Brain Res. Bull. 9 :81-86. Dahl D. and A. Bignami(1975)Glial fibrillary acidic protein from normal and gliosed human brain. Demonstration of multiple related polypeptides. Biochimica and Biophysica Acta, 386 :41-51. Dahl,D. and A Bignami(1977)Effect Of sodium dodecyl sulfate on the immunogenic prOperties of the glial fibrillary acidic protein.J. Immunological Methods, 11_:201-209. Dahl, D.,D.C. Reuger and A. Bignami(1981)Vimentin, the 57,000 molecular weight protein Of fibroblast filaments, is the major cytoskeletal component in immature glia. EurOpean Journal of Cell Biology, 24_:191-196. Dahl,D. and A. Bignami(1982)Immunohistological localization of desmin, the muscle-type 100A filament protein, in rat astrocytes and Muller glia. J. Histochem. Cytochem., 39 DeArmond, S.J., L.F.Eng,and L.J.Rubinstein(1980) The application Of glial fibrillary acidic protein immunohistochemistry in neurooncology. A progress report. Path. Res. Pract., 168 :374-394. Dixon, W. J., and M. B. Brown, eds.(1981) BMDP-81. Biomedical Computer Programs P-Series. University Of California Press, Los Angeles. Duffy,P.E.,L.Graf,and Rapport,M.M.(1978) ImmunO-cytochemical diagnoses Of brain tumor biopsies. Ann. Neurol. 1.:170. Duffy,P.E.,L. Graf, and M.M.Rapport (1979) Glial fibrillary acidic protein in ependymomas and other brain tumorszdistribution, diagnoses criteria, and relation tO differentiation. Neurol. Sci. 19 :113-146. Duffy,P.E.,M.Rapport, and L. Graf (1980) Glial fibrillary acidic protein and Alzheimer-type senile dementia.NeurOlogy, 30 :778-782. 49 Duffy, P. E., Y.-.Y. Huang, and M. Rapport (1982) The relation Of glial fibrillary acidic protein to the shape, motility, and differentiation of human astrocytoma cells. Experimental Cell Res. 189 :145-157. Freund-Mercier, M. J., and Ph. Richard(1977) Spontaneous and reflex activity Of paraventricular nucleus units in cycling and lactating rats. Brain Res. 130 : 505-520. Findley, A. L. R., J. T. Fitzsimmons, and J. Kucharczyk(1979) Dependence of spontaneously and angiotensin-induced drinking in the rat upon the oestrous cycle and ovarian hormones. J. EndocrinOl. 82 :215-225. Geisler, N., U. Plessman, and K. Weber(1982) Related amino acid sequences in neurofilaments and non-neuronal intermediate filaments. Nature. 296:448-450. Gill, J. L. (1978) Design and analysis Of experiments. Vols. 1 and 2. The Iowa State University Press, Ames. Goldman, J. E., H. H. Schaumburg, W. T. Norton(1978) Isolation and characterization Of glial filaments from human brain. J. Cell Biol. 18:426-440. Gregory, W. A., C. D. Tweedle, and G. I. Hatton (1980) Ultrasructure Of neurons in the paraventricular nucleus Of normal, dehydrated, and rehydrated rats. Brain Res. Bull. 8 :301-306. Hanakita, J., H. Fumitada, A. Shigera, E. Yamada, and H. Handa (1980) Histochemical study on oxidative enzyme activity in the brain, particularly Of astyrocytes, in spontaneously hypertensive rats. Neuropathology and Applied Neurobiology. 8 :471-482. Harris, A. B. (1973) Ultrastructure and histochemistry Of alumina in cortex. Experimental Neurology. 38_:33-63. Hatfield, J., R. P. SkOff, and L. Eng. Glial fibrillary acidic protein is localized in the lens epithelium. (1983) Submitted tO Science. Hatton, G. 1., and J. K. Walters (1973)Induced multiple nucleoli, nucleolar margination, and cell size changes in supraoptic neurons during dehydration and rehydration in the rat. Brain Res. 89 :137-154. Hatton, G. I.(1976) Nucleus Circularis: Is it an osmoreceptor in the brain? Brain Res. Bull. 1 :123-131. 50 Hatton, G. I., and C. D. Tweedle (1980) Suckling increases cell-cell apposition between neurosecretory neurons in the rat supraoptic nucleus. Soc. for Neuroscience Abstracts. 8 :457. Hatton, G. I., and C. D. Tweedle (1982) Magnocellular neuropeptidergic neurons in hypothalamus:increases in membrane apposition and number Of specialized synapses from pregnancy to lactation. Brain Res. Bull. 8 :197-204. Khachaturian, H. and J. R. Sladek (1980) Simultaneous monoamine histofluorescence and neurOpeptide immunocytochemistry :III. Ontogeny of catacholamine varicosities and neurophysin neurons in the rat supraOptic and paraventricular nuclei. Peptides. 1_: 77-95. Kimelburg, H. K., S. Narumi, and R. S. Bourke (1978) Enzymatic and morphological properties Of primary rat brain astrocyte cultures, and enzyme development 18 vivo. Brain Res. 188 :55-77. LaFarga, M., G. Palacios, and R. Perez (1975) Morphological aspects of the functional synchronization of supraOptic neurons. Experientia. 21 :348-349. Landis, M. D., and T. S. Reese (1981) Astrocyte membrane structure:changes after circulatory arrest. J. Cell Latov, N., G. Nilaver, E. A. Zimmerman, W. C. Johnson, A. J. Silverman, R. Defendini, and L. Cote (1979) Fibrillary astrocytes proliferate in response to brain inury. DevelOpmental Biology, 12 :381-384. Laursen, H., and N. H. Diemer (1980) Morphometry of astrocyte and OligOdendrocyte ultrastructure after portocaval anastomosis in the rat. Acta NeurOpathOlogy(Berl) 81 :65-70. Lazarides, E.(1980) Intermediate filament proteins as integrators of cellular space. Nature. 283 :249—256. Liem, R. (1982) Simultaneous separation and purification of neurofilament and glial filament proteins from brain. J. Neurochem. 28 :142-150. 51 Lucas, C. V., K. G. Bensch, and L. F. Eng (1980) In vitro polymerization of glial fibrillary acidic (GFA) protein extracted from multiple sclerosis (MS) brain. Neurochemical Research. 8 :247-254. Lucas, C. V., E. P. Reaven, K. G. Bensch, and L. F. Eng (1980) ImmunOperoxidase staining of glial fibrillary acidic (GFA) protein polymerized in vitro: an ultramicroscopic study. Neurochemical Research. 8:1199-1209. Manthorpe, M., R. Adler, and S. Varon (1979) Development, reactivity and GPA immunofluorescence Of astroglia-containing monolayer cultures from rat cerebrum. J. Neurocytol. 8 :605-621. Mares, V., V. Fleischmannova, Z. Lodin, and H. Ueberberg(1981) Cyclic AMP and growth regulation in rat glioma cells in tissue culture. Experimental Neurology. 11:154-160. Moonen, G., E. Heinen, and Goessens, G. (1976) Comparative ultrastructural study of the effects of serum—free medium and dibutyrl cyclic AMP on newborn brain astroblasts. Cell Tiss. Res. 181:221-227. Maxwell, D. S., and L. Kruger (1965) The fine structure Of astrocytes in the cerebral cortex and their response to focal injury produced by heavy ionizing particles. J. Cell Biol. 28;141-157. Narumi, S. H., K. Kimelburg, and R. S. Bourke (1978) Effects Of norepinephrine on the morpholgy and some enzyme activities Of primary monolayer cultures from rat brain. J. Neurochem. 21:1497-1490. Negoro, H., S. Vissessuwans, and R. C. Holland (1973) Reflex activation of paraventricular nucleus units during the reproductive cycle and in ovariectomized rats treated with oestrogen or progesterone. J. Endocrinol. §_9_ : 559-567. Nequin, L. G., J. Alvarez, and N. B. Schwartz (1979) Measurement of serum steroid and gonadotrOpin levels and uterine and ovarian variables throughout four-day and five-day estrous cycles in the rat. Biology Of Reproduction. 28:659-670. 52 Paterson, J. A., and C. P. LeBlond (1977) Increased proliferation of neuroglia and endothelia cells in the supraoptic nucleus and hypophysial neural lobe Of young rats drinking hypertonic sodium chloride solution. J. Comp. Neurol. 118:373-390. Pertti, A., H. von Kuskull, K. Teramo, O. Karjalainen, I. Virtanen, V.-P. Lehto, and D. Dahl (1980) Glial origin Of rapidly adhering amnionic fluid cells. British Medical Journal. 28:1-5. Picker, S., and S. Goldring (1982) Electrophysiological prOperties of human glia. Trends in Neuroscience. 18:73-76. Rhodes, C. H., J. I. Morrell, and D. W. Pfaff (1981) Distribution Of estrogen concentrating, neurOphysin-contaning magnocellular neurons in the rat hypothalamus as demonstrated by a technique combining steroid autoradiography and immunohistology in the same tissue. Neuroendocrinology. 33:18-23. Rueger, D. C., E. E. Gardner, H. Der Simonian, D. Dahl, and A. Bignami (1981)Purified glial fibrillary acidic protein and desmin are distinct intermediate filament proteins exhibiting similar properties.J. Biological Chemistry.‘288:10606-10612. Salm, A. K., and G. I. Hatton (1980) An immunocytochemical study Of astrocytes associated with the rat supraoptic nucleus. Soc. for Neuroscience Abstracts. 8:547. Sar, M., and W. E. Stumpf (1980) Simultaneous localization of [3-Hlestradiol and neurophysin I or arginine vasopressin in hypothalamic neurons demonstrated by a combined technique of dry-mount autoradiography and immunohistochemistry. Neuroscience Letters. 11:179-184. Sar, M., and W. E. Stumpf (1981) Combined autoradiography and immunohistchemistry for simultaneous localization of radioactively labeled steroid hormones and antibodies in the brain. J. Histochem. Cytochem. 28:161-166. Schachner, M., T. E. Hedley-Whyte, D. W. Hsu, G. Schoonmaker, and A. Bignami (1977) Ultrastructural localization Of glial fibrillary acidic protein in mouse cerebellum by immunoperoxidase labelling. J. Cell Biol. 12: 67-73. Schechter, R., C. Y. Shu-Hui, and R. D. Terry (1981) Fibrous astrocytes in senile dementia Of the Alzheimer type. J. Neuropathol. Exptl. Neurol. 18:95-101. 53 Skowsky, W. R., L. Swan, and P. Smith (1979) Effects Of sex steroid hormones on arginine vasopressin in intact and castrated male and female rats. Endocrinology. 181:105-108. Smithson, K. G., B. A. MacVicar, and G. I. Hatton (1983) Polyethylene glycol embedding: A technique compatible with immunocytochemistry, enzyme histochemistry, histofluorescence and intracellular staining. J. Neuroscience Methods. 1:27-41. Sternberger, L. A. (1974) Immunocytochemistry. Foundations Of Immunology Series. pp129-171, Prentice Hall, Inc. New Jersey. Steig, P. E., H. K. Kimelberg, J. E. Mazurkiewicz, and G. A. Banker (1980) Distribution of glial fibrillary acidic protein and fibronectin in primary astroglial cultures from rat brain. Brain Res. 199:493-500. Strocchi, P., D. Dahl, and J. M. Gilbert (1982) Studies on the biosynthesis of intermediate filament proteins in the rat CNS. J. Neurochem. 29:1132-1141. Swaab, D. F., and J. F. Jongkind (1970) The hypothalamic neurosecretory activty during the oestrus cycle, pregnancy, parturition, lactation, persistent oestrus, and after gonadectomy, in the rat. Neuroendocrinology. ‘8:133-145. Tardy, M., C. Fa es, B. Rolland, J. Bardakdjian, and P. Gonnard (1983 Effects Of prostglandins and dibutyrl cyclic AMP on the morphology of cells in primary astroglial cultures and on metabolic enzymes of GABA and glutamate metabolism.Experientia. 81:19. Theodosis, D. T., D. A. Poulain, and J.-.D. Vincent (1981) Possible morphological basis for synchronisation Of neuronal firing in the rat supraoptic nucleus during lactation. Neuroscience. 8:919—929. Trimmer, P., P. Reier, and T. Oh (1980) Morphological studies of normal and induced astrocytic differentiation .lfl vitro. Neuroscience Abstracts. 8:291. Tweedle, C. D., and G. I. Hatton (1976) Ultrastructural comparisons Of neurons of supraoptic and circularis nuclei in normal and dehydrated rats. Brain Res. Bull. 1:103-121. 54 Tweedle, C. D.,and G. I. Hatton (1977) Ultrastructural changes in rat hypothalamic neurosecretory cells and their associated glia during minimal dehydration and rehydration. Cell Tiss. Res. 181:59-72. Tweedle, C. D., and G. I. Hatton (1980a) Evidence for dynamic interactions between pituicytes and neurosecretory axons in the rat. Neuroscience. 8:661-667. Tweedle, C. D., and G. I. Hatton (1980b) Glial cell enclosure Of neurosecretory endings in the neurohypophysis of the rat. Brain Res. 192:555-559. Tweedle, C. D., and G. I. Hatton (1982) Magnocellular neurOpeptidergic terminals in neurohypOphysis: Rapid glial release of enclosed axons during parturition. Brain Res. Bull. 8:205-209. Vaughn, J. E., and Pease (1970) Electron microscopic studies Of Wallerian degeneration in rat Optic nerves. J. Comp. NeurOl. 140:207-226. Varon, S. S., and G. G. Somjen (1979) Neuron-glia interactions. Neuroscience Research Program Bulletin. #17. Virchow, R.(1840)Cellular Pathology. Trans. from the second German edition by F. Chance. London, Churchill. pp272-280. As cited in The Fine Structure Of the Nervous System. The neurons and supportin cells.A. Peters, 8. L. Palay, and H. Webster eds. (1976) w. B. Saunders CO., Philadelphia. Wakerley, J. B., and D. w. Lincoln (1973) The milk ejection reflex Of the rat: A 20-40 fold acceleration in the firing Of paraventricular neurones during oxytocin release. J. Endocrinol. 81:477-493. Watson, W. E. (1971) Some metabolic responses of rat neuroglial cells to perfusion Of the cerebral ventricles with artificial cerebrospinal fluid Of abnormal composistion. J. Physiol. 218:88-89P. Watson, W. E.(1972) A quantitative study Of some neuroglial responses to neuronal stimulation. J. Physiol.(Lond.)225 54-55P. Weber, K., and M. Osborn (1969) The reliabilty Of molecular weight determinations by dodecyl sulfate-polyacrylamide gel electrOphoresis. J. Biological Chemistry. 221:4406-4412. 55 Yarom, Y., and M. E. Spira(1982) Extracellular potassium ions mediate specific neuronal action. Science. 216:80-82. Yen, S.—H., D. Dahl, M. Schachner, and M. L. Shelanski(1976) Biochemistry Of the filaments Of brain. APPENDIX A l . APPENDIX A Protocols Immunostaining Pretreatment 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 5% hydrogen peroxide, 1/2 hour TBS rinse 2X, 10 minutes .1M sodium periodate, 15 minutes TBS rinse 3X, 15 minutes Sodium borohydride (10mg/ml), 10 minutes TBS rinse 4X, 20 minutes 5% dimethylsulfoxide (DMSO), 10 minutes TBS rinse 2X, 10 minutes 10% normal goat serum in TBS Remove normal goat serum. Briefly rinse. Immunocytochemistry Protocol 1) 2) 3) 5) 6) 7) 8) Primary anti-GFAP, 16 hours @4°c., 1:1500 in TBS TBS rinse, 1 hour Goat anti—rabbit serum, 1.5 hours @ 4°C., 1:50 in TBS Peroxidase-antiperoxidase, 1.5 hours, 1:350 in TBS TBS rinse, 4X in 2 hours Glucose-oxidase reaction, 5 minutes TBS rinse 3X All steps contain (1% triton-X 100 ’ 1 57 APPENDIX A Protocols Glucose Oxidase Reaction for Imunocytochemistry (room temperature) TO 200 mls of .l5M Phosphate Buffer pH 7.2 add: 100 mg DAB Filter solution through .22 um filter Add: 400 mg Beta-D(+) glucose 80 mg ammonium chloride .6 mg Glucose Oxidase Type VII Incubate section in this solution until brown reaction product forms. For smaller quantities of the final reaction solution, aliquots from previously mixed and frozen DAB - buffer and glucose oxidase solutions can be combined. EM-Immuno Fixative 1% glutaraldehyde 2.5% paraformaldehyde in .1M cacodylate buffer, pH 7.25 Dissolve 21.4 g Na cacodylate in 800 m1 double distilled water in a 11 beaker. Adjust pH to 7.25 Add 40 ml of stock 25% glutaraldehyde Rinse twice with distilled water 58 APPENDIX A Protocols Add 125 ml of stock 20% paraformaldehyde Transfer with three rinses to liter volumetric and Q.S. Note: HCl + formalin forms carcinogenic fumes. Pretreatment Solutions (in Tris Buffered Saline) 1) .1M Sodium periodate = 21.39g/liter = 2.139g/100mls solution 2) Sodium borohydride (lOmg/ml) = seems/semis solutionn 3) 5% DMSO = 5mls/lOOmls solution 4) 5% hydrogen peroxide = 15mls 30% H202/9Omls soln 5) 10% normal goat = 5mls serum/50mls solution TBS- Tris Buffered Saline pH 7.6 @ 25 C Measure following chemicals and transfer to a liter beaker l. 6.06 g. Tris HCL 2. 1.39 g. Tris base Adjust pH to 7.6 with appropriate acid or base solutions 3. 8.7 g. NaCl Transfer with three rinses to a liter volumetric 0.8. to the mark and stir vigorously. 59 APPENDIX A Protocols TBS-Tris Buffered Saline pH 7.6 @ 5 2 Measure the following chemicals and transfer to a 500 ml beaker 1. 3.595 g. Tris HCl 2. .250 g. Tris base Adjust pH to 7.1 with appropriate acid or base solutions 3. 4.350 g. NaCl Transfer with three rinses to a 500 ml volumetric 0.8. to the mark and stir vigorously. APPENDIX A Protocols Procedure for Thionine Staining l) Xylene lO - 15 min 2) 100% ETOH 5 min 3) 95% ETOH 2 min 4) 70% ETOH 2 min 5) Distilled water 5 min 6) Thionine stain 7) Distilled water 1 min 8) 70% ETOH 30 - 60 sec 9) 95% ETOH differentiate 10) 100% ETOH 3 min 11) Xylene 1X 5 min 12) Xylene 2X 10 min 13) Cover slip Thionine: .5% in .2M Acetate buffer, pH 5.0 61 APPENDIX A Protein purification. Protein was isolated from human postmortem cerebral white matter with the procedure described by Dahl and Bignami (1977). Briefly, tissue was extracted in .OlM sodium phosphate buffer, pH 6.2 at 4 degrees C. Further extraction was done in 0.05M phOSphate buffer. This homogenate was then centrifuged at 12,000 X g or 100,000 X g. The GFAP was then absorbed to hydroxylapitite in 0.05M sodium phosphate buffer prior to elution of GFAP with 0.01M potassium phosphate, pH 8.0. Eluates positive by immunodiffusion were pooled and precipitated with saturated ammonium sulfate (Dahl and Bignami, 1975). Further purification was achieved by extraction from sodium dodecyl sulfate polyacrylamide gels following electrOphoresis (Dahl and Bignami, 1977; Weber and Osborn, 1969)- Antiserum production and method specificity. Antibodies were produced over a period of six months by multiple backsite injections Of a total of 1mg GFAP in Freund's adjuvant. When tested with GFAP in an Ouchterlony double immunodiffusion plate, the antiserum yielded a single precipitin line. 62 APPENDIX A Protein purification—cont. Method specificity was tested on sections adjacent to those processed normally by ommission Of the primary antiserum or substitution of preimmune serum for the primary antiserum. Both procedures resulted in an abolition Of staining. Image analysis hardware/software. The imaging system consists Of 1) a TV camera, 2) Eyecom Image Scanner with a B/W TV monitor, 3) Datacolor Edge Enhancer, 4) B/W TV monitor on a 108 pt. terminal, and 5) a color TV monitor. All these components are made by Spatial Data Systems, Goleta, Ca. The terminal is interfaced with a PDP-11/34 computer. Images are transmitted to the Eyecom Scanner from photographs or negatives placed on a tungsten illuminated backlighted table. For the present project, the latter method, using black and white negatives was chosen. The image is digitized and displayed, slightly magnified, on the terminal screen consisting Of (480 X 640) 307,200 picture elements (PIXELS). The system is capable of assigning one of 256 gray level designations based upon the amount Of light transmitted through the negative to each pixel. The distribution of gray levels present in any one image can then be displayed as a frequency histogram. 63 APPENDIX A Optical equipment: 1) Zeiss microscope with a tungsten halogen light source, 2) Zeiss 10X planapo objective with anumerical aperature Of .32 and a 5.5um depth of field, 3) Zeiss achromatic, aplanatic, substage condenser, numerical aperature = 1.4, 4) #1.5 coverslips ( .16mm to .19mm thick ). The condensers were adjusted to give modified Kohler illumination in order to increase contrast and Optimize resolution. For photography, the image was directed down a mechanical tube containing a 0.5X lens to a Nikkormat camera. The Optimum wavelengh attainable with this sytem is .32um. Photography procedures: 1) Check batteries in the light meter and camera. 2) Check all lenses and glass for dirt: clean as necessary. 3) Load film. 4) Put on eyeglasses and focus Nikon eyepiece reticule. 5) Select slide and clean as necessary. 6) Find desired section through Nikkormat eyepiece, compose picture, focus. 7) Close down field diaphragm and focus it with substage condenser. 8) Open field diaphragm far enough to illuminate field as seen through eyepiece. (This was determined to be position 11 and remained constant thereafter). APPENDIX A Photography procedures-cont. 9) See a piece Of dirt on the image, curse, reclean slide and/or lenses as necessary. 10) Repeat steps 7, 8, but not 9 if lucky. 11) Move slide so that only Permount is visible thru eyepiece. Use lightmeter to adjust light to 6uA. Return slide to original position. 12) Refocus through eyepiece and reshoot picture. If not beginning a role Of film steps 5—12 were followed prior tO each shot. If beginning a role Of film additional steps were followed: (10a. Remove slide. b. Adjust light to 6uA through eye piece. 0. Place a #2 neutral density filter (1% light transmittance) on the field diaphragm and take picture. d. Repeat step C using a #0.2 neutral density filter (63% light transmittance). Thus each roll Of film had frames corresponding to 1% and 63% light transmittance. (The neutral density filters were generously loaned by Dr. James Zachs). Photomicrographs Of a stage micrometer were taken for later focussing Of the TV camera and determination Of magnification. From prior tests it was empirically determined that an ASA setting of 25 and a shutter speed of 1/250th of a second, in combination with the light level previously ' _ 65 APPENDIX A Photography procedures—cont. described and specific develOpment parameters, produced a (slightly underexposed) negative which rendered maximal detail of immunostain. Small tank film develOpment. Dilution D (1:9) Of Kodak's HC-11O develOper was used for 8 minutes at 68 degrees F with agitation every 30 seconds. Following development, stop bath (30 seconds) and then fixer (4 minutes) were added to the tank. A 15 minute wash and then application Of photo-flo finished the process. Dried negatives were mounted in a Pakon or Soligar slide binder. APPENDIX B i APPENDIX B Supplies and equipment Image Analysis Equipment Source: Spatial Data Systems P.O. Box 249 Goleta, CA 93017 prices) Combination Data Color System with Eyecom picture terminal and I/O controller for PDP/11 Refresh Memory Planes 32 Color Output Board PDP/ll Routines for console table and Camera table Density readout for model 401/704 Polaroid camera system Educational Discount (1977 $31,400.00 850.00 950.00 1,000.00 600.00 800.00 $38,900.00 1,945-00 $36,955.00 67 APPENDIX B Supplies and equipment Supply Source # Cost Tris Buffered Saline Trizma Base Sigma 500 g $14.25 Trizma Hydrocholoride Sigma 500 g 26.10 Sodium Chloride Mallinckrodt 1 lb 1.77 Triton- X 100 Mallinckrodt 500 ml 3.40 Buffer Standard Soln. pH 7.00 Mallinckrodt 1 pt 1.47 Glucose-Oxidase Reaction 3,3' —Diaminobenzidine Tetrachloride Grade II Sigma 5 g 12.50 Ammonium Chloride Mallincrodt 1 lb 1.93 B - D(+) glucose Sigma 10 g 2.50 Glucose Oxidase Type VII Sigma 10,000 units 7.15 Immunocytochemical Pretreatment Sodium meta—Periodate Sigma 25 g 4.75 Sodium Borohydride Sigma 25 g 5.30 30% Hydrogen Peroxide Mallinckrodt 1 pt 2.31 Normal goat serum Antibodies Inc. ml 2.31 Dimethyl Sulfoxide Mallinckrodt pt 5.53 37% Hydrochloric Acid Mallinckrodt 6 lb/btl 4.70 ’ . Supplies and equipment Supply 20% Aqueous EM Grade Paraformaldehyde Sodium Cacodylate Goat anti-rabbit gamma globulin P4 Normal goat serum Permount 95% EtOH Xylene 25% glutaraldehyde Gelatin Equipment Spinbar-Teflon 8 x 1.5 mm 15 x 1.5 mm 3 x 1/2 in. Latex gloves Surgeon's gloves Parafilm Weighboats 68 APPENDIX B m Polyscience Polyscience 1 pt 50 g Antibodies Inc. ml Antibodies Inc. Fisher Sci. MSU Stores MSU Stores EM Science MSU Stores 20 ml 500 ml 1 gal 5 gal 1 pt 1 lb Biochemistry Stores 1 1 1 1 1 1 roll 500 Cost $11.55 20.00 5-35 5.00 10.25 3.40 13.64 6.00 7.80 .65 6.53 18.96 APPENDIX B Supplies and equipment Equipment Pipet bulbs Test tube support Corning coverslips #1 1 cc syringes Disposable needles Falcon tubes Corning magnetic stirrer Filter System .45 um filter, triton free 13 mm filter holder 13 mm gasket .22 um filter, triton free 25 mm filter holder 25 mm gasket milli-RO4 Rogard prefilter Carbon cartridge R0 cartridge ion exchange cartridge Milli Q M2 Millipore Ft 1 pk 1 1 pk 100 100 case 100 10 100 100 12 100 Cost $2.18 3.93 5.26 12.88 8.80 63.55 85.00 34-40 23.70 15.40 31.00 47.60 31.20 1513.00 85.50 116.00 419.00 70.00 1795.00 I 7O APPENDIX B Supplies and equipment Equipment Source # Cost Pump motor 1 191.00 Reservoir 712.00 APPENDIX C ,, 71 APPENDIX C Key to ANOVA and Cell Mean Tables Type:.......Effects due to hormonal state, i.e., lactation verses estrus. Place:.....Effects due to location within the SON, i.e., anterior verses posterior or dorsal verses ventral. Pt:........Effects due to interaction of place (location) with type (hormonal state) variables. Marginal:..Means for each subgroup. Count:.....Number of subjects in group. See BMDP-81, Dixon and Brown, eds. (1981), for further reading. 72 APPENDIX C Table 3a. ANOVA table-sample size-Coronal SON DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 309.51 .0001 TYPE 1219055050.08 1 1219050505.08 .50 .5184 ERROR 9745250020.66 4 2436312505.16 PLACE 150343802.08 1 150343802.08 .05 .8297 Pt 2316991252.08 1 2316991252.08 .81 .4185 ERROR 11416441453.33 4 2854111370.83 73 APPENDIX C Table 3b. ANOVA Table-Sample size-Coronal LHA DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 1 1455.87 .0000 TYPE 22029590.08 1 22029590.08 .00 .9691 ERROR 51915515704.16 4 12978878926.16 PLACE 190937474.08 1 190937474.08 .01 .9262 Pt 154304510.08 1 154004510.08 .01 .9337 ERROR 78650833563.33 4 19662708390.83 ‘I 74 APPENDIX C Table 3c. ANOVA table-Sample size-Horizontal SON DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 1 93.14 .0024 TYPE 20878004497.06 _& 20878004497.06 2.30 .2270 ERROR 27284657193.33 3 9094885731.11 PLACE 254278272.06 1 254278272.06 .04 .8465 Pt 5396737424.06 1 5396737424.06 .94 .4028 ERROR 17145206984.33 3 5715068994.77 75 APPENDIX C Table 3d. ANOVA Table—Sample Size-Horizontal LHA DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 1 248.80 .0006 TYPE 20750272246.81 1 20750272246.81 1.45 .3143 ERROR 42801488388.58 4 14267162796.19 PLACE 2948139626.01 1 2948139626.01 4.03 .1384 Pt 6759186254.0l 1 6759186254.01 9.24 .0559 ERROR 2195362084.58 3 731787361.52 76 APPENDIX C Table 4a. ANOVA Table-Mean Density-Coronal SON DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 91823.64 1 91823.64 2103.79 .0000 TYPE 339.71 1 339.71 8.71 .0464 ERROR 167.42 4 41.85 PLACE 8.33 1 8.33 .24 .6482 Pt 17.74 1 17.74 .52 .5122 ERROR 137.44 4 34.36 77 APPENDIX C Table 4b. ANOVA Table-Mean Density-Coronal LHA DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 74127.74 1 74127.74 979.37 .0000 TYPE .18 1 .18 .00 .9632 ERROR 312.75 4 75.68 PLACE 1.24 1 1.24 .01 .9210 Pt 1.24 1 1.24 .01 .9209 ERROR 446.97 4 111.74 78 APPENDIX C Table 4c. ANOVA Table-Mean Density-Horizontal SON DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 80030.95 80030.95 54.50 .0051 TYPE 58.02 58.02 .04 .8551 ERROR 4405.62 1468.54 PLACE 11.39 11.39 .28 .6334 Pt 3.37 3.37 .08 .7921 ERROR 122.05 40.68 7 788 APPENDIX C Table 4d. ANOVA Table-Mean Density-Horizontal LHA DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 55465.83 1 55465.83 156.41 .0011 TYPE 4.18 1 4.18 .01 .9211 ERROR 1063.86 3 354.61 PLACE 18.34 1 18.34 .16 .7181 Pt 25.37 1 25.37 .22 .6725 ERROR 348.57 3 116.52 79 APPENDIX C Table 4e. Cell Means and Standard Deviations for Mean Density-Coronal LHA Cell Means CONTROL LACTATING MARGINAL ANTERIOR 71.718 79.116 78.917 POSTERIOR 78.719 77.828 78.273 MARGINAL 78.719 78.472 78.595 COUNT 3 3 6 Standard Deviations ANTERIOR POSTERIOR CONTROL LACTATING 9.219 14.477 8.211 3-557 ' , APPENDIX C Table 4f. Cell Means and Standard Deviations for Mean Density-Horizontal SON Cell Means CONTROL LACTATING MARGINAL 888818 93.266 87.164 89.605 VENTRAL 94.259 90.528 92.020 MARGINAL 93.763 88.846 90.813 COUNT 2 3 5 Standard Deviations CONTROL LACTATING DORSAL 34.611 26.575 VENTRAL 36.931 16.632 81 APPENDIX C Table 5a. ANOVA Table-Standard Deviation-Coronal SON DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. IEAN 8147.21 1 8147.21 2358.97 .0000 TYPE 11.19 11.19 3.26 .1455 ERROR 13.75 3-43 PLACE 39.56 39.56 14.25 .0195 Pt 15.57 15.57 5.61 .0770 ERROR 11.11 2.77 82 APPENDIX C Table 5b. Cell Means and Standard Deviations Deviation-Coronal SON for Standard Cell Means CONTROL LACTATING MARGINAL ANTERIOR 29.977 25.757 27.872 POSTERIOR 24.067 24.41 24.240 MARGINAL 27.022 25.090 26.058 COUNT 3 3 6 Standard Deviations CONTROL LACTATING ANTERIOR 1.317 1.160 POSTERIOR 1.457 2.661 83 APPENDIX C Table 6a. ANOVA Table-Skewness-Horizontal SON DEGREES OF MEAN *TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 1.52 1 1.52 8.90 .0584 TYPE .00 1 .00 .00 .9788 ERROR .51 3 .17 PLACE .51 1 .51 8.12 .0651 Pt .22 1 .22 5.65 .0978 ERROR .11 3 .03 APPENDIX C Table 6b. Cell Means and Standard Deviations Horizontal SON for Skewness- Cell Means CONTROL LACTATING MARGINAL DORSAL .73586 .42507 .54938 VENTRAL .06952 .36492 .24676 MARGINAL .40269 .39499 .39807 COUNT 2 3 5 Standard Deviations CONTROL LACTATING DORSAL .41752 .27905 VENTRAL .54427 .05105 85 APPENDIX C Table 60. ANOVA Table—Skewness-Horizontal LHA DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 1.88 1 1.88 9.97 .0510 TYPE .81 1 .01 .09 .7822 ERROR .57 3 -19 PLACE .30 1 .30 2.78 .1942 Pt 1.28 1 1.28 11.65 .0422 ERROR .33 3 .11 86 APPENDIX C Table 6d. Cell Means and Standard Deviations for Skewnesss Horizontal LHA Cell Means CONTROL LACTATING MARGINAL gggsgg .94742 .29972 .55880 VENTRAL -.14355 .67434 .34919 MARGINAL .40194 .48703 .45299 COUNT 2 3 5 Standard Deviations CONTROL LACTATING DORSAL .87824 .06012 VENTRAL .12820 .23259 87 APPENDIX C Table 7a. ANOVA Table-Kurtosis-Coronal SON DEGREES OF MEAN TAIL SOURCE SS FREEDOM SQUARE F PROB. MEAN 6.32 1 6.32 23.10 .0086 TYPE 1.69 1 1.69 6.20 .0675 ERROR 1.09 4 .27 PLACE 1.17 1 1.17 .54 .5021 Pt .11 1 .01 .01 .9343 ERROR 8.58 4 2.17 88 APPENDIX C Table 7b. Cell Means and Standard Deviations for Kurtosis- Coronal SON Cell Means CONTROL LACTATING MARGINAL ANTERIOR .07363 .75138 .41250 POSTERIOR .62597 1.48301 1.03949 MARGINAL .34980 1.10219 .72600 COUNT 3 3 6 Standard Deviations CONTROL LACTATING ANTERIOR .14371 1.23376 POSTERIOR .24802 1.81255 89 APPENDIX C Statistical formulas used to compute third and fourth moments about the mean. By SPSS:Skew = O = Normal < O = Platykurtic > O = Leptokurtic Kurtosis = O =Normal < O = Clustered Right > O = Clustered Left The computing formula used by SPSS is Um. no: wax-(Er Viz/J x3 {Km li’,-1)-1v.fl/(/(v Ske wness ' lhe computing fommla used by SPSS is Wg11231.”~4f(£..'t.x2)~+erz(zg,x) ..s(z;:,x,.)]/N}... _3 {[(a-nx:*)-~x1/<~-»} 90 APPENDIX C Raw Data-Anterior Coronal SON—Control Subjects gggg 13CAA 380AA 73CAA Code 13CAA 38CAA 73CAA Code 13CAA 38CAA 73CAA i£L_1§ 298724.000 203367.000 272007.000 Standard Deviation 31.506 28.858 29.566 Kurtosis .195189 .124839 -.O99146 Mean Densitx 106.174 93.776 84.585 Skewness -.118348 --351925 .275229 91 APPENDIX C Raw Data-Posterior Coronal SON-Control Subjects Code £.£l£2l§ Mean Density 13CPA 255925.00 90.485 38CPA 271444.00 92.894 73CPA 142119.00 88.860 Code Standard Deviation Skewness 13CPA 25.650 .458694 38CPA 23.771 .300636 73CPA 22.780 .351697 Code Kurtosis 13CPA .800763 38CPA .342105 73CPA .735044 92 APPENDIX C Raw Data-Anterior Coronal SON-Lactating Subjects _C_9d_e 91 CAA 32CAA 51CAA Code 91CAA 32CAA 51CAA Code 91CAA 32CAA 51CAA fi_£i£gl§ Mean Density 279763.000 81.118 188891.000 87.311 282546.000 76.886 Standard Deviation Skewness 25.859 --254250 26.878 .608423 24.563 .760567 Kurtosis 1.899755 --552935 .907311 93 APPENDIX C Raw Data-Posterior Coronal SON-Lactating Subjects Code 91CPA 32CPA 51CPA Code 91CPA 32CPA 51CPA ggd_e 91 CPA 32CPA 51 CPA fi_§l§gl§ Mean Density 290002.00 84.657 253433.00 81.762 269900.00 81.192 Standard Deviation Skewness 27.253 .619809 21.977 1.325277 24.009 .209258 Kurtosis .684895 3.523130 .150995 94 APPENDIX C Raw Data-Anterior Coronal LHA-Control Subjects Code fi_£i§gl§ Mean Density 38CAB 1353869.99 78.197 73CAB 1191411.00 69.760 13CAB 1237899.00 88.198 Code Standard Deviation Skewness 38CAB 24.063 .057023 73CAB 27.617 .535808 13CAB 31.828 .010854 Code Kurtosis 38CAB .726009 73CAB -.008867 13CAB .730821 95 APPENDIX C Raw Data-Posterior Coronal LHA-Control Subjects 9293 38CPB 73CPB 13CPB Code 38CPB 73CPB 13CPB £.£i§gl§ Mean Density 1352737.00 76.636 1186134.00 87.772 1198881.00 71.750 Standard Deviation Skewness 19.156 1.177584 21.960 .391844 26.124 .289763 Kurtosis 2.907267 1.989741 .520067 96 APPENDIX C Raw Data-Anterior Coronal LHA—Lactating Subjects Code 91CAB 32CAB 51CAB Code 91CAB 32CAB 51CAB Code 91CAB 32CAB 51CAB gm 1362686.00 1071292.00 1335836.00 Standard Deviation 21.932 30.105 26.490 Kurtosis 1.895950 -.179400 .648228 Mean Density 82-574 91.552 63.224 Skewness o936536 o729831 .310911 97 APPENDIX C Raw Data-Posterior Coronal LHA-Lactating Subjects 9993 91 CPB 32CPB 51 CPB Code 91CPB 32CPB 51CPB Code 91CPB 32CPB 51CPB fi.§i§§l§ Mean Density 1317024.00 81.235 1369870.00 78.111 1080481.00 74.137 Standard Deviation Skewness 27-385 .471655 22.621 1.326869 24.756 .027052 Kurtosis .564123 2.966640 .670475 98 APPENDIX C Raw Data-Dorsal Horizontal SON-Control Subjects Code fi_§i§gl§ Mean Density 54HDA 367258.00 117.741 66HDA 336340.00 98.387 41HDA 298178.00 56.818 Code Standard Deviation Skewness 54HDA 27.611 .440625 66HDA 22.430 .406697 41HDA 19.623 .712844 Code Kurtosis 54HDA .405781 66HDA .023060 41HDA 2.125982 99 APPENDIX C Raw Data-Ventral Horizontal SON—Control Subjects Code i Bléfilfi Mean Density 54HVA 276707.00 120.374 66HVA 350323.00 101.216 41HVA 472037.00 71.365 Code Standard Deviation Skewness 54HVA 36-387 --315331 66HAV 22.930 .366096 41HVA 23.266 .313285 Code Kurtosis 54HVA -.882780 66HVA .426432 41HVA .780499 100 APPENDIX C Raw Data- Dorsal Horizontal SON-Lactating Subjects Code 85HDA 27HDA Code 85HDA 27HDA i Pixels 191301.00 340869.00 Standard Deviation Mean Density 18-747 33.719 Kurtosis 1.864475 -.414626 68.792 106.286 Skewness 1.031089 .155657 101 APPENDIX C Raw Data-Ventral Horizontal SON-Lactating Subjects Code 85HVA 27HVA Code 85HVA 27HVA Code 85HVA 27HVA i Pixels 166426.00 264407.00 Standard Deviation Mean Density 18.638 24.039 Kurtosis -464915 .372381 68.145 99.002 Skewness .594143 -415365 102 APPENDIX C Raw Data-Dorsal Horizontal LHA-Control Subjects ‘ggde 54HDB 66HDB 41HDB Code 54HDB 66HDB 41HDB Code 54HDB 66HDB 41HDB 12.4. 688581.00 654511.00 627909.00 Standard Deviation Mean Density 14.789 21.320 29.673 Kurtosis 4-579433 .412046 --423412 85.556 84.856 49.468 Skewness 1.568427 .230460 .338521 103 APPENDIX C Raw Data—Ventral Horizontal LHA—Control Subjects Code ‘£.§i§gl§ Mean Density 54HVB 640015.00 81.413 66HVB 643848.00 87.382 41HVB 664010.00 67.454 Code Standard Deviation Skewness 54HVB 31.288 -.052896 66HVB 18.638 .594143 41HVB 21.909 .492468 .9993 Kurtosis 54HVB -.212522 66HVB .315721 41HVB 1.568397 104 APPENDIX C Raw Data-Dorsal Horizontal LHA-Lactating Subjects Code i Pixels Mean Density 85HDB 522696.00 73.790 27HDB 654239.00 91.016 Code Standard Deviation Skewness 85HDB 21.794 .326414 27HDB 28.435 .330169 Code Kurtosis 85HDB .627402 27HDB .551521 105 APPENDIX C Raw Data-Ventral Horizontal LHA-Lactating Subjects Code 85HVB 27HVB Code 85HVB 27HVB {g Pixels 395027.00 682863.00 Standard Deviation Mean Density 20.719 14.572 Kurtosis —.008766 3.941123 65.900 71.965 Skewess -.234204 -936414