Internal luminal pressure during early chick embryonic brain growthDescriptive and empirical observations.код для вставкиСкачать
THE ANATOMICAL RECORD PART A 285A:737–747 (2005) Internal Luminal Pressure During Early Chick Embryonic Brain Growth: Descriptive and Empirical Observations MARY E. DESMOND,1* MICHAEL L. LEVITAN,2 AND ANDREW R. HAAS1,3 1 Department of Biology, Villanova University, Villanova, Pennsylvania 2 Department of Mathematical Sciences, Villanova University, Villanova, Pennsylvania 3 Division of Pulmonary and Critical Care, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania ABSTRACT If the intraluminal pressure of the brain is decreased for 24 hr, the brain and neuroepithelium volumes are both reduced in half. The current study measured the intraluminal pressure throughout the period of rapid brain growth using a servo-null micropressure monitoring system. From 613 measurements made on 55 embryos, we show that the intraluminal pressure over this time period is appropriately described by a linear model with correlation coefﬁcient of 0.752. To assess whether sustained hyperpressure would increase mitosis, elevated intraluminal pressure was induced in 10 embryos for 1-hr duration via a gravity-fed drip. The mitotic density and index of the mesencephalon were measured for the 10 embryos. Those embryos, in which the colchicine solution was added to the intraluminal cerebrospinal ﬂuid creating a sustained hyperpressure, exhibited at least a 2.5-fold increase in both the mitotic density and index compared with control embryos. Based on the small sample size, we cautiously conclude that sustained hyper-intraluminal pressure does stimulate mitosis. © 2005 Wiley-Liss, Inc. Key words: CNS development; neural tube; neural tube defects; mitosis; neuroepithelium; cerebrospinal ﬂuid; CSF cell proliferation; hydrocephaly; neurogenesis The morphogenesis of the early embryonic central nervous system has fascinated embryologists for over a century. Most likely this fascination has been fueled by the fact that the central nervous system is the ﬁrst organ system to develop and features a large hollow brain at the head end of the embryo. In fact, the brain grows rapidly as a result of expansion of its cavity (Desmond and Jacobson, 1977) and coincides with completion of neural tube fusion along the dorsal midline (neurulation) and coincident occlusion of the neurocoel (Desmond and Jacobson, 1977; Schoenwolf and Desmond, 1984b, 1986; Pacheco et al., 1986; Desmond and Field, 1992). Early Embryo Brain Expansion Our laboratory, along with others, has focused on both brain expansion (Desmond and Jacobson, 1977; Desmond, 1985; Pacheco et al., 1986; Li and Desmond, 1991; Des© 2005 WILEY-LISS, INC. mond et al., 1993; Alonzo et al., 1998; Desmond and Levitan, 2002; Gato et al., 2004) and occlusion of the spinal neurocoel, which is needed to initiate the rapid expansion Grant sponsor: National Institute of Neurological Diseases and Stroke; Grant number: NINDS24136; Grant sponsor: National Institute of Child Health and Human Development; Grant number: NICHHD24710. *Correspondence to: Mary E. Desmond, Department of Biology, Mendel Science Center, Villanova University, Villanova, PA 19085. Fax: 610-519-7863. E-mail: firstname.lastname@example.org Received 27 September 2004; Accepted 11 April 2005 DOI 10.1002/ar.a.20211 Published online 23 June 2005 in Wiley InterScience (www.interscience.wiley.com). 738 DESMOND ET AL. (Desmond, 1982; Schoenwolf and Desmond, 1984a, 1984b, 1986; Desmond and Schoenwolf, 1986; Desmond and Field, 1992; Desmond et al., 1993; Desmond and Levitan, 2002). Desmond and Jacobson (1977) ﬁrst demonstrated that rapid brain expansion occurred coincident with occlusion of the spinal neurocoel. At that time, we intubated the hindbrains of living chick embryos with a small glass capillary tube for 24 hr, allowing the release of intraluminal pressure (ILP). These brains had 50% less tissue and number of cells compared to control embryos of identical stages (Hamburger and Hamilton, 1955; Desmond and Jacobson, 1977; Desmond, 1985). More recently, Inagaki et al. (1997) demonstrated a decrease in size of the posterior fossa of intubated mice embryos in utero as well. To prove that such differences were not caused by the traumatic interventive nature of the experiment, Desmond and Jacobson (1977) intubated the hindbrains of a signiﬁcant number of embryos with solid glass rods of the same outside diameter as the capillaries. The tissue volume and cell number of these sham-control embryos did not differ signiﬁcantly from the values for the nonmanipulated controls. These results suggest that the decrease in tissue volume and cell number was directly related to a decrease in pressure. To show further that brain growth depends on occlusion of the spinal neurocoel, Desmond and Levitan (2002) reported a complementary experiment. Desmond experimentally occluded the neurocoel before natural occlusion occurs and showed that the brains of these embryos expanded earlier than the brains of nonoccluded test mates. Moreover, the brains of naturally occluded embryos expanded signiﬁcantly subsequent to the natural occlusion of the neurocoel (Desmond and Levitan, 2002). Both of these experiments illustrated the dependence of brain expansion on sealing the spinal neurocoel from the outside. Both experiments point to the importance of a closed ﬂuidﬁlled system to generate a positive pressure to initiate and maintain the expansion. Embryonic ILP Measurements Intraluminal pressure measurements of the chick embryo were ﬁrst reported by Jelinek and Pexieder (1968, 1970). Using a micropipette attached to a transducer and chart recorder, and then a Landis micromanometer, these investigators proved that the pressure in the intraluminal spaces was greater than the surroundings and that there was a minimum critical pressure (approximately 2.5–3.0 mm H20) for expanding the brain vesicles. Other indirect implications of intraluminal pressure driving growth can be found in the early reports of Coulombre (1956), in which he showed that intraocular pressure is required for the normal expansion of the eye and also that intubation of the hindbrain results in folding of the roof of the hindbrain in 3-day-old chick embryos (Coulombre and Coulombre, 1958). Other investigators have demonstrated the importance of interstitial pressure in maintaining normal growth of cells and tissues, particularly in rat brain, dog lung, and frog oocytes (Lai-Fook, 1982; Wiig and Reed, 1983; Kelly and Macklem, 1991). Techniques employed for measuring interstitial ﬂuid pressure (IFP) have improved considerably over the past 60 years. One of the ﬁrst measurements of micropressure inside a living organism was made by Picken (1935) when he impaled the large protozoan, Spiratomum, with a microneedle attached to a closed cavity. This technique was followed by implanting perforated capsules subcutaneously and measuring the ﬂuid collected in the capsule via osmotic pressure (Prather et al., 1971). To overcome the limitations of the capsule technique, a wick system connected to a Statham pressure transducer was devised (Prather et al., 1971). Finally, in the 1970s, these highly traumatic and unreliable techniques were replaced by the Sanborn pressure transducer, which could measure IFP in cats and dogs that varied from ⫺6.5 to 6.0 mm Hg. The procedure was used extensively to determine the IFP in isolated dog lung (Lai-Fook, 1982; Bhattacharya et al., 1984), rat kidney (Quinn and Marsh, 1979), and rat skin and muscle (Wiig et al., 1981). Particular to the study we are reporting, Wiig and Reed (1983) used the micropressure technique to determine the interstitial ﬂuid pressure in rat brain. They demonstrated a positive correlation between IFP and the amount of cerebrospinal ﬂuid in rat brain. Goal of Current Study Using a sensitive micropressure system available from World Precision Instruments, we decided to make a critical examination of the ILP of the brain during rapid brain growth in the chick, based on the ﬁnding of Desmond and Jacobson (1977), Jelinek and Pexieder (1968, 1970), and Coulombre and Coulombre (1958), that this pressure is essential for normal brain growth. Most importantly, we were able to measure the ILP in embryos just after the time of the formation of the closed neural tube, i.e., at stage 12. Our goal was ﬁrst to establish the most appropriate mathematical model to relate ILP with the developmental stage of the brain and then to test whether application of increased pressure for a deﬁned time interval would stimulate mitosis. This hypothesis is a logical extension of our earlier ﬁndings that a decrease in ILP resulted in a signiﬁcant loss of cells comprising the neuroepithelium. Moreover, the hypothesis is consistent with earlier ﬁndings of Abercrombie (1972) that tension across cells in culture results in cell proliferation. We measured the ILP throughout the period of rapid brain growth, HH stages 12–26, with the servo-null micropressure monitoring system. Subsequently, we found that a linear model (correlation coefﬁcient of 0.752) gave an appropriate representation of the relationship between HH stages and ILP. For those embryos in which the ILP was increased for 1 hr by a constantly added colchicine infusate into the cavity, the mitotic density and index were at least 2.5 times greater when compared with the control embryos. MATERIALS AND METHODS Servo-null Controlled Micropipette System For our pressure measurements, we used the World Precision Instruments, Model 900, micropressure monitoring system (MPMS). This system is based on the fact that a change in pressure between the inside and outside of a glass micropipette will cause a change in its electrical impedance. The use of low-compliance tubing and the elimination of microscopic air bubbles make the frequency response of the hydraulic system higher than 300 Hz (Fox and Wiederhielm, 1973). A gradient of resistance vs. pressure that is sharp, reproducible, and relatively independent of the ﬁller or external salt concentration can be established if ENCEPHALIC PRESSURE DURING CHICK BRAIN GROWTH 739 Fig. 1. The ILP is shown to drop precipitously (0.5 mm Hg) when a slit is made in the roof of the midbrain. Once the initial drop occurs, the pressure somewhat stabilizes at 1.4 mm Hg. there is a relatively large difference between the electrode and the external solution (Fein, 1972). This system provides a fast, accurate, and reliable means of obtaining micropressure in biological organisms that require an interface of 5 m or less. Microelectrodes for the MPMS were made from WPI glass capillaries (o.d./i.d. ⫽ 1.20/0.68 mm with a ﬁlament) utilizing a David Kopf Instruments vertical pipette puller. The microelectrodes were backﬁlled with 3 M KCl, and the tip was beveled to a diameter of 2–5 m. Next, the microelectrode with beveled tip was placed in a microelectrode holder (WPI-MEH23SW) and electrically and hydraulically attached to the servo-pressure system. A ground wire was placed in the Petri dish holding the embryo, and ILP recordings were sent to a Macintosh MacLab chart recorder. Standardization of Micropressure Monitor To standardize the monitor system to water pressure, we measured the water pressure in mm Hg of a waterﬁlled manometer at different heights. The manometer consisted of an L-shaped glass tube (o.d./i.d. ⫽ 7/5 mm) ﬁlled with water. Once the electrodes were in position, the manometer was ﬁlled with distilled water; the entire manometer was positioned at 0° (ﬂat on the table), then raised in 10° increments from 0° to 90° against a ring stand to which a metric rule was attached. There was a linear relationship between the pressure in mm H20 and the height of the water meniscus in the column. This pressure in mm H20 was converted to the more conventionally used mm Hg. A systems check on the sensitivity of the MPMS was performed by monitoring the drop in the ILP of the brain once a slit was made in the roof of the midbrain. The pressure dropped precipitously in the ﬁrst 5-min interval, then stabilized at approximately 1.4 mm Hg (Fig. 1). Next, initial experiments were done to test whether the system would detect an increase in ILP. We noted the rise in pressure on the MPMS monitor when we blew air into the brain through a micropipette attached to plastic tubing that had been inserted into the midbrain of HH 20, 22, and 27 chick embryos (Hamilton and Hamburger, 1951). Fig. 2. The bottom chart is a typical chart of the ILP taken from the MacLab recorder. Note that there is a surge of pressure as the microelectrode is inserted into the lumen through the neuroepithelium; then the ILP stabilizes at a given pressure and returns to baseline once the electrode is removed. The top chart illustrates a typical recording of the heart beat in a stage 18 chick embryo. Note how much greater the pressure in the heart is compared to that of the embryonic brain. Since preliminary tests showed that the ILP of the embryonic brain was quite small, we assessed the functioning of the electrode by routinely monitoring the heart ILP for each embryo stage group studied. A typical QRS peak from a stage 18 embryo heart is shown in Figure 2a. Note that the ILP is approximately 100 times greater in the heart compared to what is found in the brain of the samestage embryo (Fig. 2b). Culture, Isolation, and Immobilization of Living Embryos for Monitoring Fertile White Leghorn chick eggs were incubated at 38.5°C in a forced-air humidiﬁed incubator until the embryos reached the stages of HH 12, 13, 16, 18, 20, 22, 24, and 26 (Hamburger and Hamilton, 1951). Each egg was then cracked into a ﬁnger bowl containing warm physiological saline (37°C), the vitelline membrane was removed, and the blastodisc was excised from the yolk. Next, the embryo was rinsed in physiological saline and placed 740 DESMOND ET AL. dorsal side up into a preformed depression made in wax to immobilize the embryo. The embryo was covered with prewarmed physiological saline to keep it alive and to ensure proper functioning of the microelectrode used to monitor the ILP. Measurement of ILP The ILP was obtained by impaling the midbrain an average of 10 times for at least 10 –20 sec per reading. The ILP reported is an average of these numerous impalements. No detectable loss of cerebrospinal ﬂuid (CSF) was viewed on microscopic inspection after each impalement, and the standard error (within the 95% range) of the multiple pressure measurements further supports that any loss of ﬂuid was negligible. Previous to recording the ILP for the midbrain of different stage embryos, an auxiliary experiment was done in which multiple impalements in each of the three brain vesicles were made in 12 stage 21 embryos to ensure that the pressure was constant throughout the closed neural tube based on basic principles of hydrodynamics of ﬂuids in a closed system (Gardner, 1973). We found this to be the case. For both the preliminary and actual experiment measurements, the microelectrode was connected electrically and hydraulically to the servo-pressure system, and a ground wire was placed in the physiological saline surrounding the embryo. The ILP recordings were saved on a Macintosh MacLab chart recorder. Fifty-ﬁve embryos of various stages (HH 12–26) were monitored, with the number of individual embryos per each stage varying from one to nine. Application of Sustained Hyperpressure Two dozen stage 21 chick embryos were placed in shellless culture and observed until stage 21 was attained (Hamburger and Hamilton, 1951). Elevated ILPs (10 mv) were induced into these embryos via a gravity-fed drip from a 10 ml syringe positioned 65 cm above the embryo and held by a clamp from a ring stand. Both the microelectrode and micropipette from the gravity drip tubing was placed into the midbrain. The ILP was recorded, then the solution was released into the midbrain by opening a clamp on the tubing connected to the micropipette used to impale the brain. The solution was added to the midbrain via the drip until the pressure was increased by 1 mm Hg from the initial ILP, and then the ﬂow of the drip was clamped off. A solution of physiological saline (0.9% NaCl), neutral red dye (to detect the entry of the drip solution into the brain), and colchicine (3.5 g/ml) was used as the drip solution. The purpose of the colchicine was to arrest cells of the neuroepithelium in metaphase for counting mitoses. It has long been established (Sauer, 1935) that the nuclei of the neuroepithelial cells migrate to the luminal surface when they are in the mitotic phase of the cell cycle and to the basal surface when they are in DNA synthesis. By inserting the colchicine solution into the brain cavity, the neuroepithelial cells would have been the ﬁrst to receive the mitotic arrest chemical before it had a chance to enter the systemic circulation. Elevated pressure was induced in two groups of stage 21 embryos, with ﬁve embryos per group. One group had colchicine in the infusate while another had physiological saline. The hyperpressure was created by gently releasing the drip solution once the micropipette was positioned inside the midbrain. The pressure was monitored via a microelectrode also positioned in the midbrain. The embryos were incubated throughout the duration of the experiment by submersing them in a ﬁnger bowl containing physiological saline that was maintained at 38°C in a warm room. This ﬁnger bowl was placed in a second glass bowl containing water maintained at 38°C over a hot plate. The third group of ﬁve embryos was placed into physiological saline to which 3.5 g/ml of colchicine was added. After allowing 15 min for the colchicine to enter the systemic circulation, the embryos were intubated with both the microelectrode and a glass rod having the same o.d. as the micropipette used in the chronic pressure experiments. Following the 1-hr duration of the experiments, the embryos were ﬁxed in Carnoy’s solution for 24 hr, followed by dehydration, clearing, and embedding in TissuePrep (Fisher Scientiﬁc). Serial sagittal sections were cut at 10 m, mounted onto glass slides, stained with Harris’s hematoxylin, and coverslipped in DPX. Volumes, Cell Counts, and Metaphase Assessment The areas of the embryonic brains were measured using the JAVA morphometrics software from Jandel as described in Pacheco et al. (1986). Only the midbrain was measured. Volumes were determined by multiplying the sum of the areas by the section factor and thickness of the section. The cell number was calculated by counting only nonrandomized quadrats of the midbrain using an ocular reticle. The quadrats were located in the caudal, middle, and rostral third of the midbrain. It was important to count cells in these three areas because it is well known that rates of cell proliferation differ for the regions (Kallen, 1961; Cowan et al., 1968; Wilson, 1973; Desmond and Haas, 2000). To ensure that the same cell was not counted twice, only every fourth section was used for measurements (Desmond and Haas, 2000). Once the appropriate sections were determined for the three different areas of the midbrain, the number of nuclei present in each of the small squares of the grid was counted. A minimum of 1000 cells were counted for each location. The total number of cells for the entire midbrain was extrapolated based on the relationship: number of cells/1 mm3 ⫽ number of cells/ entire midbrain volume in mm3. To determine the mitotic index and mitotic density, the number of colchicine-induced metaphases per 1,000 cells was counted using the ocular reticle. This number multiplied by 100 is the mitotic index. The mitotic density is the mitotic index per unit volume. All cell counting was done with a 40⫻ objective. RESULTS ILP and Embryonic Stages Each ILP measurement was obtained by subtracting the baseline pressure from the embryo brain pressure. The baseline pressure was actually the average of two baseline pressures: the baseline pressure before inserting the probe into the brain and the baseline pressure after removing the probe from the brain. Since the pressure measurements were extremely sensitive, this averaging took into account any variation of outside pressure during the experiment. 741 ENCEPHALIC PRESSURE DURING CHICK BRAIN GROWTH TABLE 1. Intraluminal pressure (mm Hg) in chick embryo brains HH stages Timea (hr) Embryos n Mean Standard error 95% CI 12 12.5 13 14 16 17 18 20 21 22 24 26 47 48 50 52 54 58 67 71 84 84 96 104 3 2 2 9 4 2 9 8 4 6 1 5 55b 29 15 20 116 43 22 90 94 51 63 11 59 613b 1.35 2.43 2.59 2.37 2.29 2.50 2.11 2.35 2.61 3.00 3.76 3.21 2.48c 0.102 0.114 0.210 0.071 0.126 0.115 0.083 0.039 0.046 0.053 0.241 0.088 0.031c 1.14–1.56 2.19–2.68 2.15–3.03 2.23–2.51 2.03–2.54 2.26–2.74 1.95–2.28 2.27–2.43 2.52–2.70 2.89–3.10 3.22–4.30 3.03–3.38 2.42–2.54c a Time is the average of the HH approximate interval. Sum. c For all embryos. b It was expected that the difference in baseline pressure, ⌬b, would be zero, where ⌬b ⫽ (baseline pressure after) ⫺ (baseline pressure before). However, this was not the case. It is possible that the viscosity of the intraluminal ﬂuid had an adverse effect on the sensitivity of the pressure readings. Nevertheless, as demonstrated in Figure 3, the boxplots of the data for ⌬b reveal that the average reading for each stage was very close to zero. In fact, all stages except one (stage 21) appeared to have a mean value of ⌬b not signiﬁcantly different from zero when using the t-test at the 5% level of signiﬁcance. The mean value of ⌬b for all 613 observations (across all stages) was ⫺0.032 with a standard error of 0.0140. The boxplots of the ILP are presented in Figure 4. Table 1 summarizes the data with the means, standard errors and 95% conﬁdence limits for the ILP at the various HH stages that were examined. The scatterplot of the means from Table 1 is presented in Figure 5. Note that the amount of time between embryological stages differs. This is clearly shown in Table 1. It appears that the linear model1 provides an appropriate description of the data since the analysis of variance associated with that model, ILP ⫽ 0.096 (HH stage) ⫹ 0.841, has a P ⫽ 0.005. The value of the corresponding correlation coefﬁcient was r ⫽ 0.752. This model indicates that the ILP appears to increase approximately 10% as the embryo moves from one designated HH stage to the next. HH Stage 21 Embryos With Colchicine and Pressure The second aspect of this experiment considered the impact of increased pressure on the mitotic activity of the neuroepithilium of HH stage 21 chick embryos. All of the embryos were treated with colchicine in order to assess mitosis. Speciﬁcally, there were three cases that were examined: case 1, the embryo was immersed in a colchi- 1 Various alternative models were considered for the data, such as the logistic, exponential, and logarithmic models. Using the size of the P-values associated with these analyses as the criterion for choosing a model produced no better choice than the aforementioned linear model. Furthermore, incorporating the principle of parsimony also suggested choosing the linear model. cine bath with pressure created by infused saline directly into the midbrain cavity; case 2, the CSF was withdrawn and immediately replaced by the same volume of colchicine solution with no created increase in ILP; and case 3, the ILP was increased for 1 hr (1 mm HG) by adding a colchicine infusate into the ﬂuid within the cavity. It is important to note that there was minimal collapse of the neuroepithelium in case 2, most likely because of the very short time interval (0 –30 sec) between removal of the CSF and addition of the colchicine solution. The mitotic density and the mitotic index were the aspects of mitotic activity that were analyzed in each of these three cases. It is important that caution be exercised when interpreting the results of the statistical tests due to the small sample sizes involved. Because of this, the same data were analyzed by a variety of nonparametric tests stemming from different mathematical models.2 In fact, it is only when there was consistency in the outcomes of all statistical tests applied that a suggested implication was made. The data are presented in Table 2 and the associated boxplots in Figures 6 and 7. The mitotic density is the number of mitotic ﬁgures per unit volume of tissue, while the mitotic index is the number of mitotic ﬁgures per number of cells in that tissue. Having seen no detectable intercellular substance in microscopic sections of stage 21 embryonic brains, we expected that the number of cells per unit volume of tissue for all chick embryos at stage 21 would be approximately the same. If this were true, then the mitotic density and the mitotic index essentially measure the same characteristic of mitotic activity. This is demonstrated by noting that the correlation coefﬁcients between the mitotic density and the mitotic index for cases 1, 2, and 3 are 0.966, 0.973, and 0.990, respectively. Since we were examining the potential differences among three cases for the mitotic density as well as the mitotic index, we ﬁrst employed the Kruskal-Wallis one- 2 There were insufﬁcient data to establish the likelihood of an underlying Gaussian distribution necessary for a parametric test. In general, nonparametric tests are less powerful than parametric tests. Therefore, nonparametric tests are less likely to show differences when they do exist. 742 DESMOND ET AL. Fig. 3. Boxplots of the pressure readings before and after the microelectrode was inserted into the brain. The average reading for each stage was essentially zero except for the stage 21 data. The horizontal line in the interior of each box corresponds to the median value associated with the speciﬁc data presented. The horizontal lines at the top and bottom edges of the boxes correspond to the upper and lower quartiles, respectfully. Therefore, the height of a box is the interquartile range (IQR) and contains the middle 50% of the data. A circle represents an outlier, a value 1.5 to 3 times the IQR from the adjacent upper or lower edge of the box, and an asterisk represents an extreme value, a value more than 3 times the IQR from the adjacent edge. The horizontal lines above and below the boxes represent the maximum and minimum values for the speciﬁc data set ignoring outliers and extreme values. Fig. 4. Boxplots of ILP values for all HH stage 12–26 embryos. (See explanation of boxplots in Fig. 3.) way analysis of variance test. This tested the hypothesis that the mitotic density data for each of the three cases all came from populations with the same median, and that the mitotic index data for each of the three cases all came from populations with the same median. However, the results of these tests suggested that neither the mitotic density data (exact test, P ⫽ 0.002) nor the mitotic index data (exact test, P ⫽ 0.0002) came from populations with a common median. Subsequently, we did a pairwise comparison of the medians for the various situations using the Wilcoxon-MannWhitney (WMW) test. We also examined the HodgesLehman (HL) shift estimator to determine if there appeared to be a shift between the competing pairs of distributions. We did this in two ways. First, we found a 95% conﬁdence interval for the HL shift estimator. If this conﬁdence interval contained 0, we concluded that there did not appear to be a shift in distribution; otherwise, we determined that there did appear to be such a shift. Next, we employed the permutation test based on a bootstrapping procedure of the HL shift estimator with n ⫽ 10,000. If the P-values were sufﬁciently small, we would conclude that the two treatments did appear to affect the outcomes, suggesting that there was a corresponding shift in distributions. We compared case 1 with case 2 for the mitotic density. It appeared that the WMW test (exact test, P ⫽ 0.571) suggested that the medians for these two cases were not different from one another. Furthermore, the 95% conﬁdence interval for the HL shift estimator (CI ⫽ ⫺2.66 to 6.95; in units 1 ⫻ E ⫹ 04) also suggested no shift in distributions when comparing case 1 and case 2 populations. Lastly, the permutation test for the HL shift estimator appeared to validate the same conclusion (P ⫽ 0.510). When we compared case 1 with case 2 for the mitotic index, it followed the results for the mitotic density, with the WMW test (P ⫽ 0.071), the 95% conﬁdence interval for the HL shift estimator (CI ⫽ ⫺0.76 to 6.87; in units 1 ⫻ E ⫺ 03), and the permutation test for the HL shift estimator (P ⫽ 0.076). Therefore, it appeared that there were also no differences in medians or shifts between the two distributions. Next, we compared case 1 with case 3 for the mitotic density. The WMW test (exact test, P ⫽ 0.036) suggested that the medians for these two cases were different from one another. Furthermore, the 95% conﬁdence interval for the HL shift estimator (CI ⫽ 1.40 to 3.23; in units 1 ⫻ E ⫹ 05) also suggested a shift in distributions between these populations. The permutation test for the HL shift estimator appeared to validate the same conclusion (P ⫽ 0.037). The observed values associated with the HL tests all suggest that there is a positive shift in distribution from case 1 to case 3. Once again, the results for the mitotic index followed those of the mitotic density when comparing case 1 with case 3, as we observed the WMW test (exact test, P ⫽ 0.036), the 95% conﬁdence interval for the HL shift estimator (CI ⫽ 1.06 to 2.73; in units 1 ⫻ E ⫺ 02), and the permutation test for the HL shift estimator (P ⫽ 0.038). From the HL observed values, it also appears as if there is a positive shift in distribution from case 1 to case 3. Lastly, we compared case 2 with case 3 for the mitotic density. Here, the WMW test (exact test, P ⫽ 0.008) also suggested that the medians for these two cases were different. The 95% conﬁdence interval for the HL shift estimator (CI ⫽ 1.11 to 3.03; in units 1 ⫻ E ⫹ 05) also suggested a shift in distributions between these populations, and the permutation test for the HL shift estimator appeared to validate the same conclusion (P ⫽ 0.008) as well. The observed values associated with the HL tests all suggest that there is a positive shift in distribution from case 2 to case 3. ENCEPHALIC PRESSURE DURING CHICK BRAIN GROWTH Fig. 5. 743 The scatterplot of the average ILP for each stage with its linear regression model. As before, the results for the mitotic index followed those of the mitotic density when comparing case 2 with case 3 with the WMW test (exact test, P ⫽ 0.008), the 95% conﬁdence interval for the HL shift estimator (CI ⫽ 0.72 to 2.33; in units 1 ⫻ E ⫺ 02), and the permutation test for the HL shift estimator (P ⫽ 0.008). Similar to the above, it appears that there is a positive shift in distribution from case 2 to case 3. Summarizing these results, we see that stage 21 embryo brains placed under sustained hyperpressure for 1 hr appear to have a higher mitotic density and a higher mitotic index than embryo brains whose CSF had been replaced by colchicine without pressure. When pressure was increased using saline, and the colchicine solution was only outside the embryo, the mitotic index and the mitotic density were also signiﬁcantly lower than with the experimental embryos. What makes these results signiﬁcant is the fact that they were found using less powerful statistical tests. DISCUSSION ILP With Developmental Stage It is clear from our measurements of the ILP of embryos at increasing developmental stages that a linear model is appropriate to describe the increase in ILP. Moreover, a key ﬁnding in our study is that the ILP of the embryonic brain appears to increase approximately 10% as the embryo develops from stage to stage. One extrapolation from such a ﬁnding is that it is this amount of pressure that maintains the normal turgor of the neuroepithelium during brain expansion. In fact, if the internal pressure on the neural tissue is increased by just 1 mm Hg of pressure from its normal baseline pressure for at least 1 hr, as our hyperpressure experiments reported here show, the neuroepithelium doubles its mitotic density and index. Finding that the mitotic activity of the neuroepithelium in- creased when the tension on the neuroepithelium is increased parallels the much earlier ﬁndings of Abercrombie (1970), who showed an increase in mitotic activity of cells under tension (stretch) in culture. However, ﬁnding no increase in the tissue volume in response to hyperpressure (case 3 vs. cases 1 and 2) agrees with the ﬁndings of Alonso et al. (1998) and may suggest that the cells are getting smaller after a round of cell division. Knowing that the relationship between the cavity pressure and embryonic stage is a linear one is important because it allows us to make comparisons with brain growth studies done earlier by our group and others (Desmond and Jacobson, 1977; Pacheco et al., 1986; Desmond and Levitan, 2002). We have tried to identify and understand the different mechanisms that play a role in the expansion of the cavity and the increase in volume of the neuroepithelium during early brain growth ever since our ﬁndings that a decrease in pressure via intubation resulted in a 50% reduction in the size of both. Although these recent ﬁndings suggest a role for ILP pressure, they do not exclude the role of growth factors, other proteins, or proteoglycans present in the CSF during this time period. We know from the work of Reid and Ferretti (2003) that FGF1, -2, -3, -4 proteins have been identiﬁed in the neuroepithelium of the murine choroid plexus at E12.5 da. Most likely these FGF receptors respond to FGFs in the CSF. It will be challenging to explore the interplay between growth factors and cavity pressure directing growth of the neuroepithelium. Likewise, the presence of high molecular proteins and proteoglycans in the embryonic CSF most likely plays a role in the expansion of the neuroepithelium by increasing the osmolarity of the CSF (Alonso et al., 1998; Gato et al., 2004). Proteoglycans in the embryonic CSF have been shown to originate in the neuroepithelial cells (Alonso et 744 DESMOND ET AL. TABLE 2. HH stage 21 chick embryo brains with colchicine and pressure* Case 1 (n ⫽ 3) Characteristics Cavity volume Tissue volume Total volume Cells (1 ⫻ E ⫹ 06) Mitotic ﬁgures (1 ⫻ E ⫹ 04) Mitotic density (1 ⫻ E ⫹ 05) Mitotic index (1 ⫻ E ⫺ 02) Mean Standard error 0.45 0.31 0.76 4.89 2.51 0.83 0.53 0.136 0.039 0.169 0.855 0.138 0.079 0.067 Case 2 (n ⫽ 5) Case 3 (n ⫽ 5) Mean Standard error Mean Standard error 0.76 0.35 1.10 4.23 3.32 0.98 0.81 0.075 0.031 0.103 0.454 0.323 0.114 0.087 0.66 0.33 0.99 4.44 9.01 2.82 2.13 0.087 0.030 0.107 0.537 0.558 0.293 0.267 *Case 1, embryo immersed in colchicine bath with pressure created by infused saline. Case 2, CSF withdrawn and replaced with same volume of colchicine solution for 1 hr. Case 3, CSF withdrawn and replaced with same volume of colchicine solution plus increased pressure for 1 hr. Fig. 6. Boxplots of the mitotic density for neuroepithelium of the midbrain of stage 21 chick embryos. (See explanation of boxplots in Fig. 3.) Case 3 is the true experimental case in which the intraluminal pressure was increased for 1 hr via a gravity-fed drip that contained physiological saline and colchicine. Both cases 1 and 2 are control groups for this experimental group. Case 1 had embryos that were immersed in a colchicine bath and infused with only physiological saline, whereas case 2 had embryos in which the CSF was withdrawn and replaced with the same volume of colchicine for 1 hr. There was a signiﬁcant positive shift in distribution of the mitotic density for the experimental group (case 3) relative to the control groups (cases 1 and 2). al., 1998). On the other hand, the neuroepithelial cells may simply serve as a conduit for ﬂuid from the blood plasma, as Gato et al. (2004) reported the protein composition of chick embryonal CSF to be very similar to that of blood serum for identical stages. Whatever the relationship of these chemical factors, it is important to remember that an increase in tension on the neural tissue created by an increase in internal ﬂuid alone results in an increase in mitosis. We are not excluding the role of chemical factors on neuroepithelial cell behavior; both most likely play a role. However, this experiment only tested the impact of increased tension. It is our view that both physical and chemical factors inﬂuence mitotic activity of the neuroepithelium. In contemporary developmental biology, with its emphasis on genetics and molec- Fig. 7. Boxplots of the mitotic index for neuroepithelium of the midbrain of stage 21 chick embryos. (See explanation of boxplots in Fig. 3.) The treatment for the three cases is the same as in Figure 6. There was a signiﬁcant positive shift in distribution of the mitotic index for the experimental group (case 3) relative to the control groups (cases 1 and 2). ular-cellular mechanisms, epigenetic factors such as ﬂuid dynamics are often overlooked. Recently, researchers showed the impact of intracardiac ﬂuid force on normal heart development. Embryos with severely blocked blood ﬂow failed to form several major components of the heart (Hove et al., 2003). Comparison of ILP Measurements Compared With Others’ Our ILP measurements agree remarkably with most of those reported by others who have made such measurements (Jelinek and Pexieder, 1968, 1970). For example, our ILP pressure for one stage 24, was 3.76, and they reported an average ILP of 3.8 for 38 embryos. Our measurements also show the same decrease between stages 24 and 26. Because of the close agreement of our ﬁndings with theirs, we chose to include the measurements when we only had one or two embryos in the sample. More importantly, we wanted to be able to model the trend in ILP over developmental time rather than focus intensely ENCEPHALIC PRESSURE DURING CHICK BRAIN GROWTH 745 Fig. 8. A scatterplot of the averages for the cavity and the total brain is plotted along with the averages for the ILP (data from Desmond and Jacobson, 1977; Pacheco et al., 1986). The linear model based on the ILP averages is also shown. Smooth curves are drawn through the cavity averages and through the total brain averages, respectively. These curves show how the cavity and the total brain grow in an exponential manner. One can see how, at some critical points, the increase in ILP drives the dramatic increase in the cavity and the total brain. on a particular stage of development. The only other report of ILP measured in stage 23 chick embryos differs by an order of magnitude from both ours and the JelinekPexieder group (Gato et al., 1998). Unfortunately, we are unable to make direct comparisons of our pressure ﬁndings with all of the work reported by Jelinek and Pexieder (1968, 1970; Pexieder and Jelinek, 1970) because they examined the relationship between ILP and time of incubation or embryo weight, not with embryonic stage. Most notable about our study in comparison with their ﬁndings is the fact that we veriﬁed the ILP reported by Jelinek and Pexieder (1968, 1970) using more sensitive instrumentation, and we reported ILP for earlier stages than had been measured in their studies. bolstered by our assessing the data with stringent and appropriate statistical tests. We did not determine the minimum increase in pressure necessary or the minimum amount of time that the neuroepithelium needs to be exposed to this increase in pressure for mitosis to be stimulated. Interestingly, Alonzo et al. (1998) reported that ␤-D-xyloside-treated embryos exhibited an 11% increase in the neuroepithelial volume, which was not statistically signiﬁcant using Student’s t-test. We simply showed that a sustained increase in ILP stimulated mitotic activity in the midbrain. We do know from the work of others that the rates of mitosis vary in the chick neuroepithelium for the three main vesicles and between the roof and base (Kallen, 1961; Fujita, 1962, 1963, 1964; Kallen and Valmin, 1963; Wilson, 1973). Although cell proliferation need not lead to immediate growth in the epithelium, we know that there is an increase in cell proliferation in the mesencephalon of the chick between stages 19 and 24 (Kallen, 1961). The nearly twofold increase in ILP found between stages 18 and 24 reported in our current study parallels the mitotic activity surge reported by Kallen (1961) and our own ﬁvefold increase in tissue volume for the brain (Desmond and Jacobson, 1977). Effects of Sustained Hyperpressure on Mitosis This part of the experiment was very difﬁcult to do and is the reason for the limited number of cases. Some of the difﬁculties encountered were keeping the embryo alive and warm for over 1 hr outside of an incubator and not being able to put the gravity-fed drip system inside the incubator. More importantly, it was extremely difﬁcult to maintain the increased pressure for over 1 hr. Nonetheless, after analyzing the data by a variety of nonparametric tests, we feel conﬁdent in concluding that the data support a conclusion that an increase in pressure for at least 1 hr stimulated mitosis as determined by a signiﬁcant increase in metaphases. Our conﬁdence in the data is CSF and Growth Kinetics of Early Embryonic Brain It is relevant to understanding the signiﬁcance of this work that it still remains unknown at what point the 746 DESMOND ET AL. choroid plexus begins to secrete CSF. In his early in vitro experiments, Weiss ((1934) 1955) concluded that neuroepithelial fragments placed in culture secreted ﬂuid on its ependymal surface. He had no way to illustrate that the ﬂuid came from within the cells or that it simply passed through the cells from the medium. However, knowing that growth factor receptors are present on the periventricular cells of E 12.5 mice ependyma cells, it seems probable that these same cells are able to secrete CSF. Likewise, proteoglycans in the embryonic CSF have been shown to originate in the neuroepithelial cells (Alonso et al., 1998). On the other hand, the neuroepithelial cells may simply serve as a conduit for ﬂuid from the blood plasma, as Gato et al. (2004) reported the protein composition of chick embryonal CSF to be very similar to that of blood serum for identical stages. We have shown in an earlier study that the cavity increases linearly (Pacheco et al., 1986). We assume that the cavity volume is the same as the ﬂuid volume. We have yet to discern the direct relationship between the increase in ﬂuid and the increase in pressure. When plotting the pressure measurements reported here with volume measurements reported by us in earlier studies (Desmond and Jacobson, 1977; Pacheco et al., 1986), we see that the pressure increases 50-fold between stages 18 and 24 during the same time interval that the cavity volume increases 60-fold (Fig. 8). This dramatic spurt in cavity expansion appears directly related to the burst in pressure. Prior to stage 18, the pressure seems to maintain the cavity, but there is no large increase in cavity size. It appears that stage 18 is a critical time for embryonic brain expansion and can be compared to what happens when trying to blow up a new balloon. After several trials in increasing the pressure, the walls give way and the brain immediately expands. Physiologists have explained the relationship between internal pressure and vessel expansion at any given time by stating that the distending tension in the wall of a cylindrical vessel at any time at any given pressure is directly proportional to its radius and the elastic limit of the wall (law of LaPlace after Gardner, 1973). This relationship mirrors our earlier ﬁndings that the brain exhibits spurts of growth over deﬁned time intervals. Furthermore, the expansion of the cavity and growth of the tissue alternate, i.e., ﬁrst the cavity expands faster than the tissue grows, then the tissue grows faster than the cavity expands (Pacheco et al., 1986). We interpret this alternation in growth rates as a way to ensure there is adequate tissue available to expand under the increased cavity expansion driven by the pressure increase. ACKNOWLEDGMENTS The authors thank Dr. Philip Stephens, Department of Biology, for his critical analysis of the pressure data and Mrs. Patti Sharples Haas for the tedious task of making the cell and mitotic counts. LITERATURE CITED Abercrombie M. 1970. Contact inhibition in tissue culture. In Vitro 6:128 –142. Alonso MI, Gato A, Moro JA, Barbosa E. 1998. Disruption of proteoglycans in neural tube ﬂuid by ␤-D-xyloside alters brain enlargement in chick embryos. Anat Rec 252:499 –508. Bhattacharya J, Gropper MA, Staub NC. 1984. Interstitial ﬂuid pressure gradient measured by micropuncture in excised dog lung. Am Physiol Soc 84:271–277. Coulombre AJ. 1956. The role of intraocular pressure in the development of the chick eye: I, control of eye size. J Exp Zool 133:211–225. Coulombre AJ, Coulombre JL. 1958. The role of mechanical factors in the brain morphogenesis. Anat Rec 130:289 –290. Cowan WM, Martin AH, Wenger E. 1968. Mitotic patterns in the optic tectum of the chick during normal development and after early removal of the optic vesicle. J Exp Zool 169:71–92. Desmond ME. 1982. Description of the occlusion of the lumen of the spinal cord in early human embryos. Anat Rec 204:89 –93. Desmond ME. 1985. Reduced number of brain cells in so-called neural overgrowth. Anat Rec 212:195–198. Desmond ME, Duzy MJ, Federici BD. 1993. Second messenger regulation of occlusion of the spinal neurocoel in the chick embyro. Dev Dyn 197:291–306. Desmond ME, Field MC. 1992. Evaluation of neural groove closure and subsequent initiation of spinal cord occlusion in the chick embryo. J Comp Neurol 319:246 –260. Desmond ME, Haas PA. 2000. Experimental manipulation and morphometric analysis of neural tube development. In: Tuan RS, Lo CW, editors. Methods in molecular biology, developmental biology protocols. Totowa, NJ: Humana Press. Desmond ME, Jacobson AG. 1977. Embryonic brain enlargement requires cerebrospinal ﬂuid pressure. Dev Biol 57:188 –198. Desmond ME, Levitan ML. 2002. Brain expansion in the chick embryo initiated by experimentally produced occlusion of the spinal neurocoel. Anat Rec 268:147–159. Desmond ME, Schoenwolf GC. 1986. Evaluation of the roles of intrinsic and extrinsic factors in occlusion of the spinal neurocoel during rapid brain enlargement in the chick embryo. J Embryol Exp Morph 97:25– 46. Fein H. 1972. Microdimensional pressure measurements in electrolytes. J Appl Physiol 32:560 –564. Fox JR, Wiederhielm CA. 1973. Charactistics of the servo-controlled micropipet pressure system. Microvasc Res 5:324 –335. Fujita S. 1962. Kinetics of cellular proliferation. Exp Cell Res 28:52– 60. Fujita S. 1963. The matrix cell and cytogenesis in the developing central nervous system. J Comp Neurol 120:37– 42. Fujita S. 1964. Analysis of neuron differentiation in the central nervous system by triated thymidine autoradiography. J Comp Neurol 122:311–327. Gardner WJ. 1973. The dysraphic states from syringomyelia to anacephaly. Amsterdam: Excerpta Medica. Gato A, Martin P, Alonso MI, Martin C, Pulgar MA, Moro JA. 2004. Analysis of cerebro-spinal ﬂuid protein composition in early developmental stages in chick embryos. J Exp Zool 301:280 –289. Hamburger V, Hamilton HL. 1951. A series of normal stages of the development of the chick embryo. J Morphol 88:49 –92. Hove JR, Koster RW, Forouhar AS, Acevedo-Bolton G, Fraser SE, Gharib M. 2003. Intracardiac ﬂuid forces are an essential epigenetic factor for embryonic cardiogenesis. Nature 421:172–177. Inagaki T, Schoenwolf GC, Walker ML. 1997. Change in the posterior fossa with surgically induced spina-biﬁda aperta in mouse. Pediatr Neurosurg 26:185–189. Jelinek R, Pexieder T. 1968. The pressure of encephalic ﬂuid in chick embryos between the 2nd and 6th day of incubation. Physiol Bohemoslov 17:297–305. Jelinek R, Pexieder T. 1970. Pressure of the CSF and the morphogenesis of the CNS: I, chick embryo. Folia Morphol 18:102–110. Kallen B. 1961. Studies on cell proliferation in the brain of chick embryos with special reference to the mesencephalon. Zeit Anat Entwick 122:388 – 401. Kallen B, Valmin K. 1963. DNA synthesis in the embryonic chick central nervous system. Z Zellforsch 60:491– 496. Kelly S, Macklem P. 1991. Direct measurement of intracellular pressure. Am J Physiol 260:C652–C657. Lai-Fook SJ. 1982. Perivascular interstitial ﬂuid pressure measured by micropipettes in isolated dog lung. Am Physiol Soc 82:9 –15. Li XY, Desmond ME. 1991. Modulation of Na⫹/K⫹ ATPase pumps in the heart of the chick embryo inﬂuences brain expansion. Soc Neurosci 17:2116. ENCEPHALIC PRESSURE DURING CHICK BRAIN GROWTH Pacheco MA, Marks RW, Schoenwolf GC, Desmond ME. 1986. Quantiﬁcation of the initial phases of rapid brain enlargement in the chick embryo. Am J Anat 175:403– 411. Pexieder T, Jelinek R. 1970. Pressure of the CSF and the morphogenesis of the CNS: II, pressure necessary for normal development of the brain vesicles. Folia Morphol 18:187–192. Picken LE. 1935. A note on the mechanism of salt and water balance in the Heterotrichous cilate, Spirostomum ambiguum. J Exp Biol 13:387–392. Prather JW, Bowes DN, Warrell DA, Zweifach BW. 1971. Comparison of capsule and wick techniques for measurement of interstitial ﬂuid pressure. J Appl Physiol 6:942–945. Quinn MD, Marsh DJ. 1979. Peritubular capillary control of proximal tubule reabsorption in the rat. Am J Physiol 236:F478 –F487. Reid S, Ferretti P. 2003. Differential expression of ﬁbroblast growth factor receptors in the developing murine choroid plexus. Brain Res Dev Brain Res 141:15–24. Sauer FC. 1935. Mitosis in the neural tube. J Comp Neurol 62:377– 405. 747 Schoenwolf GC, Desmond ME. 1984a. Neural tube occlusion precedes rapid brain enlargement. J Exp Zool 30:405– 407. Schoenwolf GC, Desmond ME. 1984b. Descriptive studies of occlusion and reopening of the spinal canal of the early chick embryo. Anat Rec 209:251–263. Schoenwolf GC, Desmond ME. 1986. Timing and positioning of reopening of the occluded spinal neurocoel in the chick embryo. J Comp Neurol 246:459 – 466. Weiss P. (1934) 1955. Special vertebrate organogenesis, neurogenesis. In: Willier BH, Weiss PA, Hamburger V, editors. Analysis of development. Philadelphia: Saunders. p 372. Wiig H, Reed RK, Aukland K. 1981. Micropuncture measurement of interstitial ﬂuid pressure in rat subcutis and skeletal muscle: comparison to wick-in-needle technique. Microvasc Res 21:308 –319. Wiig H, Reed RK. 1983. Rat brain interstitial ﬂuid pressure measured with micropipettes. Am Physiol Soc 83:H239 –H246. Wilson DB. 1973. Chronological changes in the cell cycle of chick neuroepithelial cells. J Embryol 29:745–751.