SEARCH

SEARCH BY CITATION

Keywords:

  • stereology;
  • disector;
  • caudal mesenteric ganglion;
  • dogs;
  • middle aging

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Aging is mostly characterized by a progressive decline of neuronal function that involves both the central and the peripheral nervous system. The aging process is accompanied by changes in either the number or the size of neurons. However, these data are controversial and not very well known in the sympathetic ganglia of large mammals. Hence, the present investigation aimed to study the dog's caudal mesenteric ganglion (CMG) in three different periods of postnatal development, searching for qualitative and quantitative alterations. The CMG is responsible for the large intestine, internal anal sphincter, and partially the urogenital system innervations. Nine dead male dogs from the Veterinary Hospital of the College of Veterinary Medicine at University of São Paulo were divided into three well-defined age groups (1–2 months old, 1–2 years old, and 5–10 years old). The stereological study was pursued using the physical disector method combined to the Cavalieri principle. The postnatal development was accompanied by an increase in the nonneuronal tissue amount and in ganglion volume. Additionally, the total number of neurons also increased during aging (from 70,140 to 1,204,516), although the neuronal density showed an opposite trend (from 29,911 to 11,500 mm−3). Due to the interrelation between either body weight or ganglion volume and aging in the dogs investigated in this study, it was possible to predict the total number of neurons in CMG using both body weight and ganglion volume in an attempt to verify whether or not size and total number of neurons are both allometrically and aging ruled, i.e., if either the animal's body weight and ganglion volume or aging influence these parameters. The prediction of the total number of neurons was very close to the initially estimated values. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

The aging process is a normal and inevitable event that occurs in all animal species. In mammals, this process is associated with decrements in cellular and physiological functions and a major incidence of degenerative diseases. The alterations that happen are the result of an interaction between many factors and aging is therefore a very complex phenomenon (Szweda et al., 2003). In the nervous system, the effects of aging are evidenced by a functional decline that involves the central and peripheral nervous system. The changes most frequently related are neuron loss, atrophy, and hypertrophy (Cabello et al., 2002). Nevertheless, these claims are discussed because there are related differences between various components of the nervous system and among animal species (Finch, 1993; Vega et al., 1993).

Our knowledge of the aging process in the sympathetic ganglia is limited, although in the last 20 years these ganglia have been studied frequently with the help of pharmacology, electrophysiology, immunohistochemistry, and biochemical techniques (Miolan and Niel, 1996). These ganglia are considered the best models to study and possibly to solve problems in the gastrointestinal tract (Gonella et al., 1987). At any time, they control important functions such as the secretion and absorption of substances from the intestinal wall and blood flux and are responsible for the arc reflex with the target tissues (Gonella et al., 1987; Bywater, 1993; Luckensmeyer and Keast, 1995, 1996; Gabella, 2004).

The degenerative alterations in the neurons of sympathetic ganglia have frequently been correlated with alterations in the dendritic and axonal arbor. However, the changes are not provoked by the direct effects of aging, but are secondary to changes in the target tissues (Andrews, 1996). With aging, tissues may decrease or increase their production of neurotrophic factors (Gavazzi and Cowen, 1996; Bennett et al., 2002; Crutcher, 2002).

The neuronal loss was often a requisite for understanding the effects of the aging process in the nervous system (West, 1994). Much research in the middle of the 1950s demonstrated a decrease in neuronal density by area (packing density) with aging in the brain, but wrongly described a decrease in the total number of neurons. Furthermore, with the development of accurate and unbiased procedures to count cells in recent years, it was possible to verify that the total number of neurons may not decrease in function with the aging process (Morrison and Hof, 1997).

Due to the lack of quantitative data concerning the cellular elements in the sympathetic ganglia in different stages of the development in large mammals and the functional importance of the caudal mesenteric ganglion (CMG) in the innervation of the gastrointestinal tract (sympathetic innervation of the colon and internal anal sphincter) and partially of the urogenital tract, this research aimed to investigate possible qualitative and quantitative alterations in the dog's CMG in two different phases of the postnatal development: maturation (pups to adult) and middle aging (adult to middle-aged).

In addition, due to the interrelation between either body weight or ganglion volume and aging in the dogs investigated in this study, it was possible to predict the total number of neurons in CMG using both body weight and ganglion volume in an attempt to verify whether or not size and total number of neurons are both allometrically and aging ruled, i.e., if either the animal's body weight and ganglion volume or aging influence these parameters.

The canine species was chosen given its importance in veterinary medicine and several gastrointestinal disorders in either middle-aged or aged dogs such as lack of motility, gastritis, and diarrhea. The causes of these diseases remain unclear and their treatment is often ineffective (Slatter, 1998; Fossum, 2002).

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Animals

In this study, we have investigated nine CMG from nine dead male dogs obtained from the Veterinary Hospital of College of Veterinary Medicine at the University of São Paulo (USP) in connection with other experiments. Animals were divided into three different groups according to their age. The age group distribution was as follows: group 1 (pups), three CMG (1 month old, 2 months old, and 2 months old; body weights: 0.15, 0.18, and 0.18 kg; cases 1, 2, and 3, respectively); group 2 (adults), three CMG (1 year old, 2 years old, and 2 years old; body weights: 13, 18, and 15 kg; cases 1, 2, and 3, respectively); group 3 (middle-aged), three CMG (5 years old, 7 years old, and 10 years old; body weights: 20, 23, and 26 kg; cases 1, 2, and 3, respectively).

Histology

After death using an overdose of anesthetic (acepromazine 0.1 mg/kg body weight injected i.v., followed by Na pentobarbitone, tiopental 50 mg/kg body weight injected i.v.), the abdominal cavity was opened by a midline incision and the intestines were slightly turned away to facilitate the identification of the ganglia and their connections to spinal cord, intermesenteric and pelvic plexuses, and also to facilitate the visualization of the main vessels of the abdomen (abdominal aorta and caudal cava vein).

About 20 ml of washing-up solution of phosphate-buffered saline (PBS; Sigma; 0.1 M; pH 7.4) containing 2% of heparin (Roche) and 0.1% sodium nitrite (Sigma) was perfused through the abdominal aorta close to the emergence of the caudal mesenteric artery. The caudal cava vein was cut to clear the circulatory system. Next, 10 ml of a 5% glutaraldehyde solution (Merck) and 1% formaldehyde (Sigma) in sodium cacodylate buffer (EMS; 0.125 M; pH 7.4) were also perfused via abdominal aorta.

Then, the ganglia were dissected out together with their connections and the caudal mesenteric artery, which passed toward the ganglia. Afterward, all ganglia were measured for their length (long axis) using a digital pachymeter (Starret) and ganglia from group 1 were fully immersed in the same fixative. Ganglia from groups 2 and 3 were sliced along their length, giving five slabs used for the volume estimate. Each slab was separately immersed in the same fixative solution in which they remained for up to 72 hs, keeping the original craniocaudal slab's orientation.

Then, ganglia (group 1) and ganglion slabs (groups 2 and 3) were washed in sodium cacodylate buffer (EMS), postfixed in 2% osmium tetroxide (EMS), block-stained with a uranyl acetate-saturated aqueous solution (Reagen), dehydrated in graded ethanols and propylene oxide (EMS), and embedded in Araldite (502 Polyscience). The resin was cured for 3 days at 60°C. Later on, 2 μm sections were cut using a glass knife, stained with toluidine blue (Nuclear), dried on a hot plate, and mounted under a coverslip with a drop of Araldite.

Morphometry

For the morphometric study, 70–80 serial sections were cut at 2 μm thickness. For each ganglion, 30 consecutive sections were collected on glass slides, stained with toluidine blue, and mounted in Araldite. A test system comprised of eight different unbiased counting frames was landed over each section field's image projected on a computer screen. A fraction (1/fr) of the counting frames was randomly, uniformly, and systematically sampled using a random start between 1 and fr. The sampled field's images were observed on a computer screen using a Leica DMR Microscope coupled with a DFC 300FX Leica digital camera. In each counting frame (with a surrounding guard area), only the neurons located inside the counting frame and not landing the forbidden lines were measured (Gundersen, 1977). This approach was also adopted in a 3D view “brick counting frame” (Howard and Reed, 2004), where the neurons from the three first sections were not measured. Each neuron received the same number in all serial sections, and its largest perikaryon profile as well as its largest nuclear profile were identified and therefore measured for the cross-sectional area using the image analysis system Q-Win Leica. The nucleus-cytoplasm ratio was also calculated by dividing the nuclear area by the cytoplasmic area. A total of 54 neurons and 33 neuronal nuclei were measured per ganglion, accounting for 162 neurons and 99 neuronal nuclei in each age group.

Stereology

A combination of the physical disector method and the Cavalieri principle was pursued to obtain an unbiased estimate of the total number of neurons (N) in the CMG (Gundersen et al., 1988, 1999; Pakkenberg and Gundersen, 1988; Mayhew and Gundersen, 1996).

Ganglion Volume (Volume Reference; Vref)

The Cavalieri principle was used to estimate the volume of the CMG. The volume is obtained using a uniform random systematic sampling. Given the small ganglion size (mean = 6.7 mm), all ganglia from group 1 were initially processed for histological study and afterwards exhaustively sectioned with a glass knife (at 2 μm thickness). In groups 2 and 3, due to the fact that the ganglia were larger (mean = 12 and 12.3 mm in groups 2 and 3, respectively), for full sectioning, all ganglia were macroscopically and transversally slabbed by means of a tissue slider before histological processing. The resulting five consecutive slabs had the same mean thickness, which was obtained by dividing the length of the ganglion by the total number of slabs using the same approach adopted by Mayhew and Olsen (1991) and Ribeiro et al. (2004). The ganglion volume was estimated by multiplying the sum of the section areas in group 1 or the sum of the sectional slab areas in groups 2 and 3 (Σa) by the distance between sections or slabs (k). The volume was estimated as Vref = k × Σa

In group 1, ganglia were exhaustively sectioned at 2 μm, which yielded an approximate total of 3,000 sections. In this approach, the interval between sampled sections (k) was 200 sections. From the first field containing ganglion tissue, the first section was sampled by taking a random number between 1 and 200. In groups 2 and 3, the mean slab thickness was 2.5 mm. In all groups, either section areas or slab areas were measured using the image analysis system Q-Win Leica. Furthermore, ganglion section's images were captured and projected on a computer screen using a DFC 300FX Leica digital camera coupled with a Leica DMR Microscope and the images of ganglion slabs were projected on a computer by means of a TK 1280OU JVC camera coupled with an Axioscopic Microscope Zeiss (L08). The area of the caudal mesenteric artery was not considered in the estimate of either section or the ganglion slab area.

The accuracy of the volume estimate was evaluated through the coefficient of error of the Cavalieri principle (Gundersen and Jensen, 1987). The formula used to assess the error coefficient (CE) was CE = 1/ΣA × [1/12 (3a + c − 4b)]1/2, where ΣA is the sum of all section or slab areas; a is the sum of all products a × a; b is the sum of all products a × (a + 1); and c is the sum of all products a × (a + 2).

The tissue shrinkage effects were calculated using the following formula: shrinkage = volume before processing − volume after processing/volume before processing. The tissue shrinkage using Araldite embedding was 15%. Therefore, the volume of the reference space estimated using the Cavalieri principle was corrected as follows: Vref (corrected) = V(ref) × (1 − shrinkage) (Braendgaard et al., 1990; Howard and Reed, 2004).

Neuronal Density (Nv)

From the serial 2 μm pairs of sections (disectors), each one consisted of a reference and a look-up section with a fixed height (h) of 10 μm, were selected throughout the extent of the ganglia. A test system comprised of eight different unbiased counting frames was landed over the image of each section field on a computer screen. A fraction (1/fr) of counting frames was randomly, uniformly, and systematically sampled using a random start between 1 and fr. The sampled field's images were observed on a computer screen. The total sampled area (a) in the test systems was 10,000 μm2 for group 1 and 40,000 μm2 for groups 2 and 3. Neuronal profiles that appeared in a specific area of the reference section but did not appear in the same area in the look-up section were counted, being referred to as Q. Then, the neuronal density was calculated by the sum of Q in all disectors divided by the sum of all disector volumes, i.e., the product between the area test (a) and the height (h) of the disector (V(dis) = a × h). Nv = ΣQΣV(dis).

The sampling scheme employed was based on the results of a pilot study in which about 34 disectors were considered in group 1 and 20 disectors in groups 2 and 3 (4 disectors being applied in each ganglion slab). The interval (k) between each disector was 100 sections in group 1 and 250 sections in groups 2 and 3. The first reference section was randomly chosen between 1 − k, being a section between 1 and 100 in group 1 and between 1 and 250 in groups 2 and 3. Each ganglion from group 1 yielded 30 disectors, whereas each ganglion from groups 2 and 3 gave us 24 disectors.

The reference and look-up sections were projected on a computer screen. The reference section was drawn up on a transparency and then compared to the image of the look-up section seen on the computer screen.

Total Number of Neurons (N)

The total number of neurons was obtained as the product of the ganglion volume (Vref) and the numerical density (Nv) following the equation: N = Vref (corrected) × Nv

Volume Density (Vv)

The same reference sections used to calculate the numerical density were considered for the volume density estimate. Volume density represents the fraction of total CMG volume occupied by neurons. It is obtained by randomly throwing a point grid system over the reference sections. Next, the total number of points falling within the reference space was counted (P(rs)). Also, the total number of points landing in cell bodies was counted (P(cb)). Then, the Vv was estimated as: Vv = ΣP(cb)/ΣP(rs). The results were expressed as a percentage (Howard and Reed, 2004).

Mean Neuronal Volume (Vn)

From estimates of volume density and numerical density, the mean neuronal volume was obtained (Mayhew, 1989). It was calculated as the ratio between Vv and Nv: Vn = Vv/Nv.

Statistical Analysis

Morphometric and stereological data were analyzed using the Statistical Analysis System (SAS, 1995). The analysis of variance (ANOVA) was performed between each age group in all morphometric and stereological parameters to assess the effect of age, especially on the total number of neurons (N). When the results were considered significant (P < 0.05), the Tukey test was pursued.

The correlation analysis between all stereological parameters and body weight was carried out using Pearson's product moment. Correlation can be classified as higher intensity (0.7–1.0), medium intensity (0.5–0.7), and lower intensity (0.1–0.5).

The stereological results were also analyzed through a (logistic-based) regression model (linear regression) to test the interrelation between body weight and all stereological parameters, i.e., numerical density, ganglion volume, total number of neurons, volume density, and mean neuronal volume, and also the interrelation between the ganglion volume and the same former parameters. From linear functions, Y = a + by,x · X (where Y is a dependent variable, a is the intercept of the regression line with the x-axis, by,x represents the slope of regression line, and x is the independent variable), it was possible to predict the values of the stereological parameters plotted against either body weight or ganglion volume. The straight correlation between both predicted and estimated values was verified through the determination coefficient (R2), which ranged from 0 to 1. When R2 is higher (almost 1), the function is adequate and by all means both predicted and estimated values are very close to one another. Conversely, when R2 is lower (< 0.5), the function is inadequate, meaning that there is no correlation between estimated and predicted values.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Macrostructural Organization

The CMG was located close to the abdominal aorta involving the emergence of the caudal mesenteric artery. The dorsal portion of the ganglion was connected to the sympathetic trunk by lumbar splanchnic nerves and its cranial portion to the celiac-mesenteric ganglion through intermesenteric nerves. Three different nerves left the caudal pole of the ganglion: the main hypogastric and the left and the right hypogastric nerves. The main hypogastric nerve followed the caudal mesenteric artery to the colon. The left and the right hypogastric nerves were directed caudally toward the pelvic cavity reaching the pelvic ganglia.

The main hypogastric nerve divided into two separated branches following the two branches of the caudal mesenteric artery. The former branch running cranially was called the left colonic nerve, whereas the second, running caudally, was named the cranial rectal nerve. The cranial rectal and left colonic nerves gave rise to branches to the descending colon and rectum.

In all groups analyzed, a single ganglionic mass was macroscopically seen. The ganglion length ranged from 6.0 to 7.0 mm in pups (group 1; 6.7 ± 0.06), it was constant in adult dogs (group 2) at 12.0 mm, and ranged from 10.0 to 16.0 mm in middle-aged dogs (group 3; 12.3 ± 0.32).

Microstructural Organization

Under the light microscopy, the structure of the CMG was remarkably different, especially between pups (group 1) and adult or middle-aged dogs (groups 2 and 3). In pups (group 1), it was observed that the CMG was not constituted of a single ganglionic mass surrounded by a capsule as in adults (group 2) and middle-aged dogs (group 3). However, it was formed by 2–4 ganglionic masses (or ganglionic lobes) surrounded individually by a capsule that contained connective tissue and vessels (Fig. 1).

thumbnail image

Figure 1. Microstructure of the dog's CMG in a pup showing the typical arrangement in three ganglionic lobes or ganglionic masses (1, 2, and 3) surrounded individually by a capsule that contained connective tissue and vessels. Toluidine blue. Scale bar = 30 μm.

Download figure to PowerPoint

In adults (group 2) and middle-aged dogs (group 3), the capsule sent septa of connective tissue inside the CMG, dividing it into ganglionic units. Each ganglionic unit was composed of various cell types and the units were separated from each other by nerve fibers, intraganglionic capillaries, and septa of collagen fibers. Given this cytoarchitectural arrangement, the CMG was described as a true ganglionic complex. The ganglion capsule thickness ranged from 11.8 to 21.3 μm (15.2 ± 3.6) in pups, from 17.6 to 21.3 μm (19.9 ± 2.0) in adults, and from 16.7 to 28.3 μm (23.5 ± 5.7) in middle-aged dogs.

The main cell types observed in the CMG were ganglion neurons, glial cells, and SIF cells. Ganglion neurons were generally spindle-shaped and readily distinguishable due to their large size, clear nucleus, and the evident nucleolus. The nucleus was predominantly eccentric. In pups, it was observed that the nucleolus was not a single structure, but in fact two nucleoli were frequently observed. Furthermore, it was found that 3% of neurons were binucleate in pups. In middle-aged dogs, an aging pigment, i.e., lipofucsin granules, was distributed toward the cytoplasm.

Ganglion neurons were surrounded by a thin glial capsule containing one to three glial cell nuclei around the neuron (Figs. 2–4). Moreover, SIF cells were seen arranged in two different ways, i.e., close to neurons and encompassing tight clusters comprised of 2–3 cells in the proximity of blood vessels.

thumbnail image

Figure 2. Micrographs of a pair of sections (2 μm thick and stained with toluidine blue) used for estimating nerve cell numbers in pup (group 1). In each pair, the reference section is at the top (A). B is the look-up section. The separation between reference section and look-up section is 10 μm. Black arrows show two neuronal transects seen in the reference section, which no longer exist in the look-up section (white arrows). Scale bar = 30 μm.

Download figure to PowerPoint

thumbnail image

Figure 3. Micrographs of a pair of sections (2 μm thick and stained with toluidine blue) used for estimating nerve cell numbers in an adult dog (group 2). In each pair, the reference section is at the top (A). B is the look-up section. The separation between reference section and look-up section is 10 μm. Black arrows show three neuronal transects seen in the reference section, which no longer exists in the look-up section (white arrows). In A and B, large white arrowheads show two glial cell nuclei. Scale bar = 30 μm.

Download figure to PowerPoint

thumbnail image

Figure 4. Micrographs of a pair of sections (2 μm thick and stained with toluidine blue) used for estimating nerve cell numbers in an aged dog (group 3). In each pair, the reference section is at the top (A). B is the look-up section. The separation between reference section and look-up section is 10 μm. The black arrow shows one neuronal transect seen in the reference section, which no longer exists in the look-up section (white arrow). Scale bar = 30 μm.

Download figure to PowerPoint

Morphometric Study

Neuronal cross-sectional area.

The cross-sectional area of 162 neurons in each age group was calculated. It ranged from 125.5 to 1,035.5 μm2 (435.0 ± 177.0) in group 1 (pups), from 109.1 to 1,966.1 μm2 (980.2 ± 359.5) in group 2 (adults), and from 100.0 to 3,539.1 μm2 (1,185.8 ± 569.8) in group 3 (middle-aged dogs; Fig. 5).

thumbnail image

Figure 5. Histograms documenting the distribution of neuronal sizes (cross-sectional area of the largest profile of a neuron). The top histogram shows the percentage distribution of sizes divided into classes of the same size for the three age groups and ranging from the smallest (100–400 μm2) to the largest (3,400–3,700 μm2). The other three histograms illustrate a single age group each and present the distribution of cell body sizes in 11 classes evenly spread between the maximum and minimum values for those age groups.

Download figure to PowerPoint

In group 1, the majority of neurons had a cross-sectional area varying from 100 to 400 μm2 (45.7%), which was close to the cross-sectional area frequency observed from 400 to 700 μm2 (43.2%). In groups 2 and 3, the majority of neurons, 37% and 27.3%, respectively, presented a cross-sectional area sized from 700 to 1,000 μm2 or from 1,000 to 1,300 μm2, particularly for group 3. In the middle-aged dogs group, a small percentage of neurons (9.3%) varied from 1,900 to 3,600 μm2, which are the largest neurons recorded (Fig. 5).

Nuclear cross-sectional area.

The cross-sectional area of 99 neuronal nuclei was calculated in each age group. It varied from 63.5 to 178.5 μm2 (115.2 ± 24.7) in group 1, from 85.3 to 297.3 μm2 (179.6 ± 41.3) in group 2, and from 49.7 to 417.7 μm2 (187.3 ± 70.66) in group 3 (Fig. 6).

thumbnail image

Figure 6. Histograms documenting the distribution of nuclear sizes (cross-sectional area of the largest profile of a neuron nucleus). The top histogram shows the percentage distribution of sizes divided into classes of the same size for the three age groups and ranging from the smallest (50–85 μm2) to the largest (400–435 μm2). The other three histograms illustrate a single age group each and present the distribution of neuron nuclei sizes in 11 classes evenly spread between the maximum and minimum values for those age groups.

Download figure to PowerPoint

The majority of the nuclei in group 1 had a cross-ssectional area sized between 85 and 120 μm2 (42.4%). In groups 2 and 3, the majority of nuclei, 37.4% and 33.4%, respectively, had a cross-sectional area ranging from 155 to 190 μm2. In the middle-aged dog group, a small percentage of neurons (11.1%) presented a nuclear area varying from 295 to 435 μm2, which were the largest nuclei measured (Fig. 6).

Nucleus-cytoplasm ratio.

The nucleus-cytoplasm ratio (nuclear cross-sectional area/cytoplasm cross-sectional area) was verified in 99 neurons in each age group. It ranged from 0.14 to 1.53 (0.41 ± 0.22) in group 1, from 0.11 to 1.90 (0.23 ± 0.18) in group 2, and from 0.04 to 0.66 (0.2 ± 0.08) in group 3.

In group 1, 29.3% of neurons had a nucleus-cytoplasm ratio, varying from 0.34 to 0.44. In group 2, the majority of neurons presented a ratio class between 0.14 and 0.24 (60.6%). Finally, in group 3, 58.6% of neurons showed a nucleus-cytoplasm ratio class between 0.14 and 0.24 (Fig. 7).

thumbnail image

Figure 7. Distribution of nucleus-cytoplasm ratio classes and the relative frequencies in all age groups (pup, adult, and aged). The classes were evenly spread in the same size (0.10).

Download figure to PowerPoint

Stereological Study

Ganglion volume (volume reference; Vref).

The ganglion volume, estimated through the Cavalieri principle, ranged from 1.9 to 2.5 mm3 (mean = 2.3; CV = 0.14) in group 1 (pup), from 78 to 94.5 mm3 (mean = 87.2; CV = 0.09) in group 2 (adult), and from 93.5 to 124 mm3 (mean = 104.5; CV = 0.16) in group 3 (middle-aged). The error coefficient for the volume estimate was 1.8%, 2%, and 3% in group 1; 5.3%, 7.6%, and 7.6% in group 2; and 7%, 10%, and 9% in group 3.

Numerical density (Nv).

The numerical density estimated using the physical disector method ranged from 29,143 to 32,258 mm−3 (mean = 29,911; CV = 0.07) in group 1 (pup), from 12,500 to 13,158 (mean = 12,719; CV = 0.03) in group 2 (adult), and from 10,978 to 11,818 (mean = 11,500; CV = 0.04) in group 3 (middle-aged; Figs. 2–4).

Total number of neurons (N).

The total number of neurons, estimated using the physical disector method associated with the Cavalieri principle, ranged from 55,448 to 81,677 (mean = 70,140; CV = 0.19) in group 1 (pup), from 975,000 to 1,243,421 (mean = 1,110,307; CV = 0.12) in group 2 (adult), and from 1,026,468 to 1,452,534 (mean = 1,204,516; CV = 0.18) in group 3 (middle-aged).

Volume density (Vv).

The volume density ranged from 32.5% to 39% (mean = 35.3%; CV = 0.09) in group 1 (pup), from 27.1% to 31.8% (mean = 29.1%; CV = 0.08) in group 2 (adult), and from 28.8% to 30.6% (mean = 29.4%; CV = 0.03) in group 3 (middle-aged).

Mean neuronal volume (Vn).

The mean neuronal volume ranged from 11,166 to 12,131 μm3 (mean = 11,800; CV = 0.04) in group 1 (pup), from 21,680 to 24,183 μm3 (mean = 22,867; CV = 0.06) in group 2 (adult), and from 24,488 to 26,261 μm3 (mean = 25,667; CV = 0.04) in group 3 (middle-aged). All the stereological parameters investigated in this study are summarized in Table 1.

Table 1. Overall view of all stereological parameters investigated in the CMG from dogs at three distinct ages
GROUPAGEBW1Nv2Vref3N4Vv5Vn6
  • 1

    Body weight (Kg).

  • 2

    Numerical density (mm−3).

  • 3

    Ganglion volume (mm3).

  • 4

    Total number of neurons.

  • 5

    Volume density (%).

  • 6

    Mean neuronal volume (μm3).

I (pup)1 month old0.1529,1432.573,29432.511,166
 2 months old0.1828,3331.955,44834.412,131
 2 months old0.1832,2582.581,6773912,093
Mean 0.1729,9112.370,14035.311,800
CV 0.10.070.140.190.090.04
II (adult)1 year old1312,500891,112,50028.422,720
 2 years old1813,15894.51,243,42131.824,183
 2 years old1512,50078975,00027.121,680
Mean 15.3312,71987.21,110,30729.122,867
CV 0.160.030.090.120.080.06
III (middle-aged)5 years old2010,97893.51,026,46828.826,261
 7 years old2311,7041241,452,53430.626,152
 10 years old2611,818961,134,54528.924,488
Mean 23.0011,500104.51,204,51629.425,667
CV 0.130.040.160.180.030.04

Coefficient of Correlation and Analysis of Regression

In the CMG, the correlation between animal body weight and all stereological parameters was tested by means of Pearson's product moment correlation coefficient and analysis of regression. The body weight of animals, total number of neurons, ganglionic volume, and the mean particle volume demonstrated a positive correlation to each other. However, when these parameters were plotted against the numerical density and volume density, a negative correlation was observed. The correlation among stereological parameters was classified as high intensity except for volume density, which presented a medium intensity correlation.

Given the high intensity correlation described for body weight and ganglion volume when plotted against further stereological parameters, with the exception of volume density, a linear regression analysis, expressed by the function Y = a + by,x · X, was pursued in an attempt to verify the interrelationship between predicted and estimated stereological parameter values (Figs. 8 and 9). Hence, body weight was assumed to be a fixed value (X) and all other parameters (Y) were therefore plotted. By the same token, the ganglion volume was also pointed out as a fixed value. By all means, the predicted values were fairly close to the estimated ones, as evidenced by the high determination coefficient (R2), except for volume density, which showed a low determination coefficient when body weight was assumed to be a fixed value (R2 = 0.55), and when ganglion volume was considered to be a fixed value (R2 = 0.53).

thumbnail image

Figure 8. Graphs representing the linear functions between the body weight (kg) and the numerical density (Nv), ganglion volume (Vref), total number of neurons (N), and mean neuronal volume (Vn). The values were represented by triangles for estimated values and diamonds for predicted values. Notice that the plots do not contain values for pups.

Download figure to PowerPoint

thumbnail image

Figure 9. Graphs representing the linear functions between the ganglion volume (Vref) and the numerical density (Nv), total number of neurons (N), and mean neuronal volume (Vn). The values were represented by triangles for estimated values and diamonds for predicted values. Notice that the plots do not contain values for pups.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

Technical Approach

For the quantitative work, it is absolutely necessary to see and define clearly the counting unit and therefore avoid shrinkage. Thus, the choice of the adequate fixative, i.e., a modified Karnovsky solution delivered by vascular perfusion, and a plastic resin embedding were crucial for the development of this experiment. Due to shrinkage problems caused by paraffin embedding, a plastic resin such as Araldite is preferred instead (Guillery and Herrup, 1997; Von Bartheld, 2002; Gardella et al., 2003) and also the Araldite-embedded block sectioning is a requirement for both measurements and quantification, which provides an outstanding resolution and sharpness.

In this study, the cell body rather than the nuclei or nucleolus was considered the counting unit. The prevertebral ganglia of mammalian comprise not only mononucleate neurons as reported by Miolan and Niel (1996) but also binucleate neurons, which have been found in the guinea pig (Szurszewski and Miller, 1994; Ermilov et al., 2000), capybaras (Ribeiro et al., 2004), and rabbit (Sasahara et al., 2003). Given the reasons above, the choice of the nucleus as a counting unit was not feasible. Furthermore, the literature shows that binucleate neurons can be observed in the initial phases of development (Szurszewski and King, 1989; Appenzeller, 1990) and at this stage it is unknown if in the dogs from group 1 (1–2 months old) the cell division has been completed. By the same token, the use of the nucleolus for quantification was not possible either because of the presence of one or two nucleoli as reported for the celiac-mesenteric ganglion (Miller et al., 1996; Ribeiro et al., 2002).

Morphological Aspects: Macro- and Microstructure

The location of the dog's CMG was the same in the different age groups as reported by Goshal (1986), Evans (1993), and Gagliardo et al. (2003). Macroscopically, the CMG was seen as a unique ganglion mass, as also reported by Gagliardo et al. (2003). However, Goshal (1986) has reported that the CMG can be divided into two ganglion masses in the dogs and four ganglion masses in cats. Although more than one ganglion mass was not macroscopically observed, it was microscopically verified that the pups' CMG comprised two to four ganglion masses (or ganglionic lobes), which is in agreement with the observations made by Goshal (1986).

The microstructural arrangement of the dog's CMG was similar between adult and middle-aged dogs, e.g., the CMG was divided into distinct compartments by capsular septa of connective tissue as reported for mammalian prevertebral ganglia by Gabella et al. (1988), Szurszewski and King (1989), Banks (1992), Szurszewski and Miller (1994), Miolan and Niel (1996), Schmidt (1996), and Gagliardo et al. (2003). In pups, however, two or four ganglion masses were noticed and they were structurally linked by connective tissue and each one was ensheathed by a ganglion capsule. The ganglion capsule thickness had a 30% increase from pup to adult, a 18% increase from adult to middle-aged, and a 54.6% increase from pup to middle-aged animals.

Morphometric Study

The size of the CMG neurons, expressed as their cross-sectional area, showed a 2.25-fold increase from pup to adult and 1.2-fold from adult to aged, which was significant in both cases (P < 0.01). The results of the present study are in agreement with those reported for the celiac-mesenteric and superior cervical ganglion of rats (Baker and Santer, 1988) and of humans (Schmidt, 1996), for the rat's distal vagal ganglion (Soltanpour et al., 1996), and for the rat's hypogastric ganglia (Warburton and Santer, 1997).

The increase in size of neurons between pups and adults can be explained by the incomplete maturation of those at the early stages of development as reported by Vega et al. (1993) and Masliukov (2001). However, the significant neuronal increase between adult and middle-aged is susceptible to some speculation such as the possibility of the perikaryon enlargement given the accumulation of lipofucsin granules (Finch, 1993; Warburton and Santer, 1997), due to the reduction of the pressure exercised by the surrounded cells as a result of cell death (Baker and Santer, 1988), or as a compensation mechanism for the neuronal loss with development (Finch, 1993; Warburton and Santer, 1997).

Not only an increase of the cell body but also in the nuclear size was observed during the development. However, the nuclear increase was significant (P < 0.01) only when pups were compared to other groups, a 1.56-fold increase being verified from pups to adults and a 1.04-fold increase from adults to middle-aged animals. The nuclear increase during maturation (from pups to adults) was already expected since this phase is characterized by an increase of nuclear volume (McMahon et al., 2003).

The nucleus-cytoplasm ratio of the CMG neurons decreases with aging, as stated by Ledda et al. (2000). This ratio decreases when the area occupied by the cytoplasm in either adults or aged is larger than the area occupied in pups, or when the area occupied by the nucleus in either adults or aged is smaller than those occupied in pups, or by means of the association of these two factors. In this study, it has been verified that although both nucleus and cytoplasm increased in size, the cytoplasm increase was proportionally larger than that of the nucleus during aging.

Stereological Study

Ganglion volume.

The increase in both sensory and parasympathetic ganglion volume during development has already been reported in the literature, such as in the rat's distal vagal ganglion (Soltanpour et al., 1996), rat's dorsal root ganglion (Popken and Farel, 1997), rat's hypogastric ganglion (Warburton and Santer, 1997), and cervical (C5) and lumbar (L4) (Bergman and Ulfhake, 1998).

Similar results were now found in a sympathetic ganglion, i.e., the dog's CMG, where a 37.9-fold increase was shown from pups to adults and a 45.4-fold increase from pups to middle-aged animals. These figures were significant (P < 0.01), though any increase between adult and a middle-aged animal was of no significance.

Numerical density.

The numerical density in the CMG of pups was higher and significant (P < 0.01) when compared to the additional groups. In adults and in middle-aged dogs, the numerical density was 42% and 38%, respectively, of that found in pups. Yet, though a decrease was found between adults and middle-aged animals, these results were of no significance.

It should also be stressed the lower coefficients of variation (CV) for the numerical density (Nv) within age groups may express a tight control of the interindividual and age-related variation in the density of neurons in a certain volume of the CMG. That may imply a functional role in the target organ innervation as well as in the synaptic organization within CMG. Coefficients of variation were of 7% in pups, 3% in adults, and 4% in middle-aged dogs.

The biological meaning of CV was stated by Clegg (1983), who performed morphometric studies of the spleen of hypoxic mice. In that experiment, coefficients of variation of spleen variables showed a tendency to decrease in animals exposed to an atmospheric pressure of 72 kPa, but rose markedly at 52 kPa. This finding indicated that at 72 kPa spleen and red pulp changes are adaptive, but at 52 kPa they indicated an overall relative failure of adaptive mechanisms, with consequent reduced somatic fitness. Hence, when things are biologically important for survival or function, there is often a tight control (Clegg, 1983).

An age-related decrease in neuronal density was also observed in dorsal root ganglia, demonstrating a reduction of 50% in neuronal density when 2- to 3-month rats were compared to 30-month rats (Popken and Farel, 1997) and also a 40% decrease in hypogastric ganglia's neuronal density when 4-month rats were compared to 24-month rats (Warburton and Santer, 1997).

The high numerical density in pups was due to the small size of neurons and also the volume density was higher. In addition, with aging, a 17% increase of nonneuronal tissue amount (connective tissue, vessels, and neuropil) was reported in our present investigation and, in line, a reduction of approximately 2.35-fold and 2.6-fold in the numerical density was observed when pups were compared to both adults and middle-aged animals, respectively. Yet the aging process was accompanied by a 17% decrease in the neuronal volume density when pups were compared to both adults and middle-aged animals. Further, the mean neuronal volume increased significantly (P < 0.01) with aging when all groups were compared to one another.

Total number of neurons.

Although it is easy to think that the phases of development are associated with a decrease in the number of neurons, despite some controversy about this dogma (Crutcher, 2002), and that apoptosis is a natural process of development, they are no more than a good deduction (Finch, 1993). The results of this study demonstrated that in the dog's CMG there was no reduction in the total number of neurons from adult to middle-aged dogs. Moreover, the number of neurons found in adult and middle-aged dogs is larger and more significant (P < 0.01) than the total number of neurons estimated in a puppy.

The number of neurons in the CMG of pups represented only 6.3% of that quantity in adults and 5.8% of that in middle-aged dogs. The small percentage of neurons obtained in pups in relation to the other groups could be associated with the small ganglion volume in pups when compared to the other age groups.

In addition, a nonsignificant 40% increase in the total number of neurons in the hypogastric ganglion of rats of 24 months compared to rats of 4 months was reported by Warburton and Santer (1997) and either a 28% significant increase (using the physical disector method) or a 19% significant increase (using the point counting method) in the total number of neurons in the dorsal root ganglia of rats of 80 days compared to rats of 11 days was reported by Popken and Farel (1997). It was stated by the latter that the increase in the number of neurons could be associated to possible cell division in different developmental periods.

However, Farel (2003) reported that the increase in the number of neurons in the dorsal root ganglia of rats is not associated with a possible neurogenesis, but may be the result of a later maturation or incomplete differentiation, which makes it difficult to identify the neurons that will not be quantified due to the fact that they do not show a typical morphology.

By contrast, a reduction in the number of neurons with aging was shown in the superior cervical ganglion, where rats of 4 months presented a not significantly larger number of neurons than rats of 24 months. In that study, the physical fractionator method was pursued (Santer, 1991). Similar results were also reported by Bergman and Ulfhake (1998), who observed a reduction in the number of neurons in the dorsal root ganglia of rats of 3 months when compared to rats of 30 days using the optical disector.

Hence, the 17-fold increase in the total number of neurons of the dog's CMG might be related to three main factors: neuronal division, later maturation or incomplete neuronal differentiation, and binucleate neuron division. In the literature, there is no robust evidence for cell division, at least not for the rat DRG (Popken and Farel, 1997). Instead, late maturation or incomplete neuronal differentiation seems to be the case for the rat's hypogastric ganglion (Farel, 2003).

As for the third possibility, i.e., binucleate neuron division, it seems very unlikely that a 3% population of binucleate neurons as seen in pups (which would account for a 6% increment after the mitotic division) would be responsible for a 17-fold increase in CMG neurons as reported in this investigation. Nevertheless, cell markers (Activin, BrdU) should be tested for large mammals' sympathetic ganglia in the near future.

Relationship Between All Stereological Parameters and Body Weight or Ganglion Volume

Due to the interrelation between either body weight or ganglion volume and aging in the dogs investigated in this study, it was possible to predict the total number of neurons in CMG using both body weight and ganglion volume in an attempt to verify whether or not size and total number of neurons are both allometrically and aging ruled, i.e., if either the animal's body weight and ganglion volume or aging influence these parameters.

From the linear regression analysis, it was possible to predict the interrelationship between body weight and all stereological parameters, i.e., total number of neurons, neuronal density, ganglion volume, and mean neuronal volume, as well as the relationship between ganglion volume and the following stereological parameters: total number of neurons, neuronal density, and mean neuronal volume. It has been shown that the results estimated in this study allow a close correlation to the predicted results, verified by the high determination coefficient (R2). A similar study was pursued by Mayhew (1991) using the relationship between cerebellar weight and the number of Purkinje cells of different animal species. However, it was carried out using a different stereological method than was used in our study, i.e., the fractionator. The results obtained reflected a positive correlation between number and Purkinje layer surface area within a given species; further, the surface area was related to the body weight of each species.

Although our study shows a close correlation between both predicted and estimated results (stereological parameters) and either body weight or ganglion volume, the results using the ganglion volume allow a better correlation with the estimated results. Although some of the predicted results slightly corresponded to the estimated values, the reliability of the function expressed by the determination coefficient (R2) was high for all parameters, with the exception of the volume density when plotted against the body weight and ganglion volume, where only 55% and 53% correlations, respectively, were verified between predicted and estimated results.

From the linear functions obtained, it was possible to predict that dogs of 180 g, 15 kg, and 23 kg would have in the CMG, respectively, 106,041, 982,111, and 1,455,023 neurons; 28,945, 15,083, and 7,600 neurons/mm3; a volume of 4.3, 81.1, and 122.5 mm3; a volume density of 34%, 30%, and 28%; and a mean neuronal volume of 12,126, 22,262, and 27,733 μm3. In our study, the estimated results for the same dogs were 81,677, 1,243,421, and 1,452,534 neurons; 32,258, 13,158, and 11,704 neurons/mm3; a volume of 2.5, 94.5, and 124 mm3; a volume density of 39%, 31.8%, and 30.6%; and a mean neuronal volume of 12,093, 24,183, and 26,152 μm3.

By the same token, a similar prediction was possible from the CMG volume. For instance, a dog with a ganglion volume of 2.5, 94.5, and 124 mm3 would have, respectively, 77,015, 1,139,524, and 1,480,219 neurons; 29,203, 12,688, and 7,393 neurons/mm3; a volume density of 35%, 30%, and 28%; and a mean neuronal volume of 11,985, 23,989, and 27,838 μm3. The estimated results for the same dogs were 81,677, 1,243,421, and 1,452,534 neurons; 32,258, 13,158, and 11,704 neurons/mm3; a volume density of 39%, 31.8%, and 30.6%; and a mean particle volume of 12,093, 24,183, and 26,152 μm3. The correlation between the stereological parameters and body weight or between stereological parameters and the ganglion volume can be easily shown in the above example, where an increase in body weight is accompanied by an increase in ganglion volume and consequently by an increase in the total number of neurons.

On the other hand, an increase in both ganglion volume and body weight was accompanied by a decrease in the number of neurons per volume unit (mm−3; numerical density) inside the ganglion. Moreover, an increase in the mean neuronal volume is correlated with an increase in both body weight and ganglion volume. In this way, it is possible to show that a dog of 180 g has either a ganglion volume or a total number of neurons smaller than those in dogs of 15 and 23 kg. Conversely, both numerical density and volume density in dogs of 180 g are larger than in dogs of 15 and 23 kg.

In conclusion, this study sheds light on some aspects of postnatal development in the dog's CMG, which was represented here by the period from pups to middle-aged dogs. The main aspects seen here were the total number of neurons, CMG's microstruture, neuronal density, and neuron size.

Although there is still controversy as to what happens to the total number of neurons in aging (specially in large mammals) and in different parts of the nervous system (enteric nervous system and central nervous system), the results presented in this investigation, which were obtained by using modern and current unbiased stereological methods (Baddeley, 2001) for number estimation, revealed a nonreduction in this parameter during aging, which increased instead. Furthermore, both body weight and ganglion volume are positively correlated to aging in dogs and either body weight or ganglion volume allows an accurate prediction of the total number of neurons in the dog's caudal mesenteric ganglion during the postnatal period.

In addition, due to the wide territory of innervation of CMG in dogs, further studies (using neurotracers) may focus on specific CMG's cell populations in order to figure out what happens, for instance, to the total number of rectum-innervating neurons or descending colon-innervating neurons during aging in an attempt to elucidate the causes of several gastrointestinal disorders especially in middle-aged or aged dogs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED

A.A.C.M.R. thanks Professor Giorgio Gabella (University College London, U.K.) for his advice with the preparation of the material (fixation, embedding, sectioning) and cytological interpretation of the results during aging, Professor Terry M. Mayhew (University of Nottingham, U.K.) for advice, help with the use of the physical disector, and the linear regression analysis interpretation performed in this study, as well as Emerson Ticona Fioretto for his advice with the preparation of the images that illustrate this study. K.M.G. thanks Naianne Kelly Clebis for her assistance in computer-assisted image acquisition.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. LITERATURE CITED
  • Andrews TJ. 1996. Autonomic nervous system as a model of neuronal aging: the role of target tissue and neurotrophic factors. Microsc Res Tech 35: 219.
  • Appenzeller O. 1990. Anatomy and histology. In: AppenzellerO, editor. The autonomic nervous system. Oxford: Elsevier. p 110.
  • Baddeley A. 2001. Is stereology “unbiased”? Trends Neurosci 24: 375376.
  • Baker DM, Santer RM. 1988. Morphometric studies on pre- and paravertebral sympathetic neurons in the rat: changes with the age. Mech Aging Dev 42: 139145.
  • Banks WJ. 1992. Sistema nervoso. In: BanksWJ, editor. Histologia veterinária aplicada. São Paulo: Editora Manole. p 318351.
  • Bennett MR, Gibson WG, Lemon G. 2002. Neuronal cell death, nerve growth factor and neurotrophic models: 50 years on. Auton Neurosci 95: 123.
  • Bergman E, Ulfhake B. 1998. Loss of primary sensory neurons in the very old rat: neuron number estimates using the dissector method and confocal optical sectioning. J Comp Neurol 396: 211222.
  • Braendgaard H, Evans SM, Howard CV, Gundersen HJG. 1990. The total number of neurons in the human neocortex unbiasedly estimated using optical disectors. J Microsc 157: 285304.
  • Bywater RAR. 1993. Activity following colonic distension in enteric sensory fibres projecting to the inferior mesenteric ganglion in the guinea-pig. J Auton Nerv Syst 46: 1926.
  • Cabello CR, Thune JJ, Pakkenberg H, Pakkenberg B. 2002. Aging of substantia nigra in humans: cell loss may be compensated by hypertrophy. Neuropathol Appl Neurobiol 28: 283291.
  • Clegg EJ. 1983. Morphometric studies of the spleen of the hypoxic mouse. J Microsc 131: 155161.
  • Crutcher KA. 2002. Aging and neuronal plasticity: lessons from a model. Auton Neurosci 96: 2532.
  • Ermilov LG, Miller SM, Schmalz PF, Hanani M, Szurszewski JH. 2000. The three-dimensional structure of neurons in the guinea pig inferior mesenteric and pelvic hypogastric ganglia. Auton Neurosci 83: 116126.
  • Evans HE. 1993. The autonomic nervous system. In: EvansHE, editor. Miller's anatomy of the dog. London: W.B. Saunders. p 776799.
  • Farel PB. 2003. Late differentiation contributes to the apparent increase in sensory neuron number in juvenile rat. Dev Brain Res 144: 9198.
  • Finch CE. 1993. Neuron atrophy during aging: programmed or sporadic? Trends Neurosci 6: 104110.
  • Fossum T. 2002. Cirurgia de pequenos animais. São Paulo: Editora Roca.
  • Gabella G, Trigg P, Mcphail H. 1988. Quantitative cytology neurons and satellite glial cells in the superior cervical ganglion of the sheep: relationship with ganglion neuron size. Neurocytology 17: 753769.
  • Gabella G. 2004. Autonomic nervous system. In: GabellaG, editor. The rat nervous system. London: Academic Press. p 81103.
  • Gagliardo KM, Silva RA, Guidi WL, Ribeiro AACM. 2003. Gross and semi-thin view of the caudal mesenteric complex of the domestic dog (Canis familiaris—Linnaeus, 1758). Anat Histol Embryol 32: 18.
  • Gardella D, Hatton WJ, Rind HB, Rosen GD, Von Bartheld CS. 2003. Differential tissue shrinkage and compression in the z-axis: implications for optical dissector counting in vibratome-, plastic- and cryosections. J Neurosci Methods 124: 4559.
  • Gavazzi I, Cowen T. 1996. Can the neurotrophic hypothesis explain degeneration and loss of plasticity in mature and aging autonomic nerves? J Auton Nerv Syst 58: 110.
  • Gonella J, Bouvier M, Blanquet F. 1987. Extrinsic nervous control of the motility of small and large intestines and related spincters. Physiol Rev 67: 902961.
  • Goshal NG. 1986. Inervação abdominal, pélvica e caudal autônoma. In: GettyR, editor. Anatomia dos animais domésticos. Rio de Janeiro: Guanabara Koogan. p 16281634.
  • Guillery RW, Herrup K. 1997. Quantification without pontification: choosing a method for counting objects in sectioned tissues. J Comp Neurol 386: 27.
  • Gundersen HJG. 1977. Notes on the estimation of the numerical density of arbitrary profiles: edge effect. J Microsc 111: 219223.
  • Gundersen HJG, Jensen EB. 1987. The efficiency of systematic sampling in stereology and its prediction. J Microsc 147: 229263.
  • Gundersen HJG, Bagger P, Bendtsen TF, Evans SM, Korbo L, Marcussen N, Moller A, Nielsen K, Nyengaard JR, Pakkenberg B, Sorensen FB, Vesterby A, West MJ. 1988. The new stereological tools: disector, fractionator, nucleator and point sampled intercepts and their use in pathological research and diagnosis. Acta Pathol Microbiol Immunol Scand 96: 857881.
  • Gundersen HJG, Jensen EB, Kiâu K, Nielsen J. 1999. The efficiency of systematic sampling in stereology: reconsidered. J Microsc 193: 199211.
  • Howard CV, Reed MG. 2004. Three-dimensional measurement in microscopy. In: HowardCV, ReedMG, editor. Unbiased stereology. Oxford: Bios Scientif Publishers.
  • Ledda M, Barni L, Altieri L, Pannese E. 2000. Decrease in the nucleo-cytoplasmic volume ratio of rabbit spinal ganglion neurons with age. Neurosci Lett 286: 171174.
  • Luckensmeyer GB, Keast JR. 1995. Distribution and morphological characterization of viscerofugal projections from the large intestine to the inferior mesenteric and pelvic ganglia of the male rat. Neuroscience 66: 663671.
  • Luckensmeyer GB, Keast JR. 1996. Immunohistochemical characterisation of viscerofugal neurons projecting to the inferior mesenteric and major pelvic ganglia in the male rat. J Auton Nerv Syst 61: 616.
  • Masliukov MP. 2001. Sympathetic neurons of the cat stellate ganglion in postnatal ontogenesis: morphometric analysis. Auton Neurosci 89: 4853.
  • Mayhew TM. 1989. Stereological studies on rat spinal neurons during postnatal development: estimates of mean perikaryal and nuclear volumes free from assumptions about shape. J Anat 162: 97109.
  • Mayhew TM. 1991. Accurate prediction of purking cell number from cerebellar weight can be achieved with the fracionator. J Comp Neurol 308: 162168.
  • Mayhew TM, Olsen DR. 1991. Magnetic resonance imaging (MRI) and model-free estimates of brain volume determined using the cavalieri principle. J Anat 178: 133144.
  • Mayhew TM, Gundersen HJG. 1996. “If you assume, you can make an ass out of u and me”: a decade of the dissector for stereological counting of particles in 3D space. J Anat 188: 115.
  • McMahon SS, Dockery P, Mcdermott KW. 2003. Estimation of nuclear volume as an indicator of maturation of glial precursor cells in the developing rat spinal cord: a stereological approach. J Anat 203: 339344.
  • Miller SM, Hanani M, Kuntz SM, Schmalz PF, Szurszewski JH. 1996. Light, electron and confocal microscopic study of the mouse superior mesenteric ganglion. J Comp Neurol 365: 427444.
  • Miolan J, Niel J. 1996. The mammalian sympathetic prevertebral ganglia: integrative properties and role in the nervous control of digestive tract motility. J Auton Nerv System 58: 125138.
  • Morrison JH, Hof PR. 1997. Life and death of neurons in the aging brain. Science 278: 412418.
  • Pakkenberg B, Gundersen HJG. 1988. Total number of neurons and glial cells in human brain nuclei estimated by disector and fractionator. J Microsc 150: 120.
  • Popken GJ, Farel PB. 1997. Sensory neuron number in the neonatal and adult rats estimated by means of the stereologic and profile-based methods. J Comp Neurol 386: 815.
  • Ribeiro AACM, Elias CF, Liberti EA, Guidi WL, De Souza RR. 2002. Structure and ultrastructure of the celiac mesenteric ganglion complex in the domestic dog. Anat Histol Embryol 31: 344349.
  • Ribeiro AACM, Davis C, Gabella G. 2004. Estimate of size and total number of neurons in superior cervical ganglion of rat, capybara and horse. Anat Embryol 208: 367380.
  • Santer RM. 2001. Sympathetic neuron numbers in ganglia of young and aged rats. J Auton Nerv Syst 33: 221222.
  • SAS. 1995. User's guide: basic and statistic. Cary, NC: SAS Institute. p 1686.
  • Sasahara THC, Souza RR, Machado MRF, Silva RA, Guidi WL, Ribeiro AACM. 2003. Macro- and microstructural organization of the rabbit's celiac-mesenteric ganglion complex (Oryctolagus cuniculus). Ann Anat 185: 441448.
  • Schmidt RE. 1996. Neuropathology of human sympathetic autonomic ganglia. Microsc Res Tech 35: 107121.
  • Slatter D. 1998. Sistema gastrointestinal. In: SlatterD, editor. Manual de cirurgia de pequenos animais. São Paulo: Editora Manole. p 591760.
  • Soltanpour N, Baker DM, Santer RM. 1996. Neurons and microvessels of the nodose (vagal sensory) ganglion in young adult and aged rats: morphometric and enzyme histochemical studies. Tissues Cell 28: 593602.
  • Szurszewski JH, King BF. 1989. Physiology of prevertebral ganglia in mammals with special reference to inferior mesenteric ganglion. In: SchultzSG, WoodJD, RaunerBB, editors. Handbook of gastrointestinal physiology. Bethesda, MD: American Physiological Society. p 519577.
  • Szurszewski JH, Miller SM. 1994. Physiology of the prevertebral ganglia. In: JohnsonLR, editor. Physiology of the gastrointestinal tract. New York: Raven Press. p 795878.
  • Szweda PA, Camouse M, Lundberg KC, Oberley TD, Szweda LI. 2003. Aging, lipofuscin formation, and free radical-mediated inhibition of cellular proteolytic systems. Aging Res Rev 2: 383405.
  • Vega JA, Calzada B, Del Valle ME. 1993. Age-related in the mammalian autonomic and sensory ganglia. In: AmentaF, editor. Aging of the autonomic nervous system. London: CRC Press. p 3161.
  • Von Bartheld CS. 2002. Counting particles in tissue sections: choices of methods and importance of calibration to minize biases. Histol Histopathol 17: 639648.
  • Warburton AL, Santer RM. 1997. The hypogastric ant thirteenth thoracic ganglia of the rat: effects of age on the neurons and their extracellular environment. J Anat 190: 115124.
  • West MJ. 1994. Advances in the study of age-related neuron loss. Sem Neurosci 6: 403411.