Effects of In Utero Phrenic Nerve Section on the Development of Collagen and Elastin in Lamb Lungs 1- 3
OWEN S. BAMFORD, AWILDA RIVERA, TERESA TADALAN, and WILLIAM ELLIS Introduction
I f a fetus does not breathe normally in
utero, lung development is delayed, and lungs at birth are smaller and less compliant than normal lungs. Although abnormalities of lung morphometry have been documented in such lungs (1, 2), there is little information on other factors that could contribute to their abnormal mechanical properties. The purpose ofthis study was to examine whether the abnormal mechanical properties are associated with alterations in the deposition of the extracellular structural proteins collagen and elastin. Many workers have examined the effects of experimental abolition or reduction of fetal breathing movements on lung development. Methods used have included bilateral phrenic nerve section (1-3), cordotomy (4), imitation diaphragmatic hernia (5), thoracoplasty to increase thoracic compliance (4), and oligohydramnios (6). Tracheostomy has also been used to bypass the upper airway resistance and therefore to abolish the small distending pressure in the fetal lungs (7). In all these studies, the lungs of the experimental group were smaller and less compliant than those of the control group. Histologic studies of human infant lungs with phrenic nerve agenesis (8), fetal lamb lungs after phrenic nerve section (1), and fetal guinea pig lungs after oligohydramnios (9) all showed severe lung hypoplasia with reduced air space, and similar findings occurred after fetal tracheostomy (7). Thus, there is a consensus that interference with fetal breathing, however caused, affects morphometric and mechanical properties of the lungs. The mechanisms by which fetal breathing movements affect lung properties are less clear. Surfactant does not appear to be involved; surfactant production was not affected by tracheostomy or phrenic nerve section (2, 3, 7). Another possible mechanism is via lung connective tissue. 1202
SUMMARY Interference with fetal breathing movements Is known to retard morphologic development of the lung and to reduce compliance. We hypothesized that the lower compliance might be In part due to effects on lung structural proteins. We studied the effects of phrenic nerve section in utero on lung compliance and on the lung contents of collagen, elastin, and DNA. At 110to 112 days of gestation, one fetal lamb In each of 12 twin pregnancies had either both phrenic nerves cut (PX) or a sham operation (S). The other twin was left unoperated (Upx, Us) as a control. They were killed 14to 22 days later, and the concentrations In lung parenchyma of collagen (as hydroxyproline HPro), elastin, and DNA were measured, together with lung compliance and dry and wet weight. Paired comparisons were made (PX versus Upx and S versus Us). Both operated groups (PX, S) had smaller lungs with lower water content than did their unoperated twins. Absolute static compliance In PX was reduced, but compliance relative to lung weight was unchanged, and there was no significant difference between S and Us. There were no significant effects of PX on the concentrations of HPro, elastin, and DNA, or on the elastin/collagen ratio. Compliance was not correlated with either HPro or elastin content. HPro content Increased significantly with gestational age in all groups. It Is concluded that phrenic nerve section retards the Increase of lung compliance and possibly air space, but It does not affect the overall rate of lung cell proliferation or of deposition of elastin or collagen. Fetal breathing appears to promote physical but not biochemical aspects AM REV RESPIR DIS 1992; 146:1202-1205 of lung development.
The mechanical properties of the lung depend partly on its connective tissue development (10, 11). Schellenberg and coworkers (12) found that both collagen and elastin content of fetal lamb lungs increased during late gestation as compliance increased, and Noguchi and colleagues (13) showed a developmental increase in elastin mRNA in late-gestation bovine lungs. Fetal lamb lungs treated to accelerate lung maturation showed increased lung parenchymal elastin and collagen as well as greater compliance (14). Hypoplastic lungs from human fetuses appear to have less collagen IV in their alveolar septa than do control lungs (15). Thus the increase in compliance might in part be due to increased production of these structural proteins. Conversely, the abolition of fetal breathing could affect lung compliance by reducing their production. In this study we tested the hypotheses that (1) fetuses whose breathing movements are abolished by phrenic nerve section have parenchymal collagen and elastin contents that are reduced compared with those of their unoperated twins, and (2) that the tissue content of one or both of these proteins is cor-
related with lung compliance on air inflation. Methods Surgery Fourteen pregnant Dorset-cross ewes were operated at 110to 112 days of gestation (term, 138 to 140 days), using standard aseptic methods. In the 12twin pregnancies, surgery was performed on only one fetus, the other being left intact as a control. Anesthesia was induced intravenously by 1g sodium pentothal and maintained by ventilation with 20/0 halothane in oxygen. The uterus was approached by a midline incision and opened for access to a fetus, which was withdrawn as far as the umbilicus. The thorax was opened laterally usually between the ninth and tenth ribs, and (Receivedin originalform November 15, 1991and in revised form May 7, 1992) 1 From the Divisionof Neonatology, Department of Pediatrics, University of Maryland, Baltimore, Maryland. 2 Supported by a ResearchGrant from the American Lung Association of Maryland. 3 Correspondence and requests for reprints should be addressed to Dr. O. S. Bamford, Johns Hopkins Children's Center, Eudowood Division of Pediatric Respiratory Sciences, Park 316, 600 N. Wolfe St., Baltimore, MD 21205.
1203
FETAL BREATHING AND WNG STRUCTURAL PROTEINS IN LAMBS
the ribs and lung were retracted separately to expose the inferior vena cava with the phrenic nerve attached. The nerve was lifted and cut at the point where it leaves the vena cava, central to its division over the tendinous portion of the diaphragm. Care was taken to avoid the lymphatic trunks. A pair of stainless steel wire electrodes were sewn into the diaphragm. The ribs and intercostal muscles were then closed by a single layer of sutures, and the skin was closed as a separate layer. The procedure was repeated on the other side. Catheters were placed in the fetal carotid artery and trachea, and recording electrodes wereplaced for electrocortical activity (BCoG) and nuchal extensor electromyogram (BMG). The uterus, body wall, and skin were closed. Sham-operated preparations were treated identically except that the phrenic nerves were identified and lifted but not cut.
Monitoring Fetal arterial pressure and heart rate, ECoG, nuchal and diaphragm EMG, and tracheal pressure were recorded continuously on an eight-channel chart recorder (Grass Instruments, Quincy, MA) until the animals were killed. The diaphragm EMG and tracheal pressure wererecorded to confirm absence of breathing movements. Fetal arterial blood samples were taken daily to check that blood gas and pH values were in the normal range. Measurements At 14 to 22 days postsurgery, preparations were ended by maternal injection of T-61 euthanasia solution (20 ml intravenously). Placental passage of this agent causes deep fetal anesthesia, with loss of ECoG activity within seconds of injection, and cardiac arrest occurs after 1 to 2 min. Compliance. The pressure-volume relationship in static air inflation was used to calculate static lung compliance. Both operated and unoperated fetuses wereimmediately removed from the uterus and weighed, and the thorax was opened carefully to expose the lungs. No attempt was made to remove fluid from the lungs, as it was felt that any such attempt would introduce another uncontrolled variable. A cannula was tied into the trachea, and the lungs were inflated with air from a syringe in 1-mlsteps. To control for the dynamic response of the lung to step-volumechanges, pressure was recorded at a standard time of 30 s after the inflation step, and the next 1ml was immediately added. The process continued until either 10 ml had been added or until the lungs leaked. If there was no leak at 10 ml inflation then air was withdrawn in 1-mlsteps to zero,with pressure measurements continuing as before. Pressure-volume curves were drawn for operated and unoperated lungs, and the volumes at 10 and 20 em H 2 0 were determined graphically (VlO and V20). Only VlO was used for analysis as there was a strong correlation between the two volumes (r = 0.974) and because in one PX and three S fetuses the lungs leaked at 20 em H 2 0 . The phrenic nerves and diaphragm were examined
for evidence of reinnervation. (This was found in only one PX preparation, which showed apparently normal breathing after surgery and was reassigned to the S group.) The lungs were removed, weighed, and frozen for later analysis. Dry weight. Five samples were taken from each pair of lungs (approximately 1 g each sample), one from each right upper, lower, and middle lobe, and one each from the left upper and lower lobe. Samples were weighed wet and then dried at 95° C to constant weight and reweighed to give dry weight. Collagen content. Collagen was estimated from hydroxyproline content of acid hydrolysate. Five tissue samples from the areas of lung used for dry weight wereminced together. Five samples (approximately 100mg each sample) were taken from the mince and hydrolyzed separately in 4 ml6N HCl for 16 h at 100° C. Then 100-lllaliquots of each hydrolysate were dried in vacuum and redissolved in water for hydroxyproline assay by oxidation with chloramine-T at room temperature for 20 min, followed by heating to 60° C for 15 min with Ehrlich reagent. The optical density was then read at 550 nm (16). The assay was calibrated with standard hydroxyproline solutions; the curve was linear over the concentration range found. As a check on the colorimetric hydroxyproline assay, aliquots were taken from 34 hydrolysates from sevenfetuses, dried in vacuum, and redissolved for assay on a Beckman amino-acid analyzer (Beckman Instruments, Fullerton, CA). The correlation coefficient r between determinations by the two methods was 0.94. Elastin content. Insoluble elastin was assayedgravimetrically by the Lansing hot-alkali method, as recommended by Soskel and coworkers (17) for lung tissue. Three 1-gmixed samples were taken from each pair of lungs, minced, and suspended in 30 ml O.lN NaOH in 60 ml polypropylene centrifuge tubes. The tubes were heated in a boiling water bath for 45 min, then cooled and centrifuged for 10 min at 3,000 rpm, and the supernate was poured off. The precipitate was resuspended in cold O.lN NaOH and centrifuged a second time. The supernate was poured off, and the precipitate was washed with cold O.lN NaOH onto an accurately preweighedWhatman GP5 glass fiber disk (Whatman, Clifton, NJ) in a vacuum filtration apparatus. The precipitate was washed twice by filtration with cold O.1N NaOH and then twice more in cold deionized water. Finally, the disk was removed from the filtration manifold, dried in a 60° C oven to constant weight, and weighed on a microbalance accurate to 0.1 mg (Mettler, Hightstown, NJ). DNA content. DNA was measured as an indication of total cell number. Three mixed l-g samples werehomogenized (Cole-Parmer, Chicago, IL) in deionized water to form a 20070 suspension, which was then treated to extract DNA (18). Protein and nucleic acids were precipitated with 10% trichloroacetic acid (TeA) at room temperature; the precipitate was washed with 95% ethanol twice and then
suspended in 5% TeA at 90° C. The protein precipitate was washed twice with 5% TeA, the supernate fractions were combined, and the absorbance was read at 260 nm for DNA concentration.
Statistical Treatment Fetal lambs vary greatly in body weight and developmental stage at both surgery and death. Data on lung development show very large individual variation, and unpaired comparisons between group means can require a prohibitive number of subjects to demonstrate statistical differences. To reduce this source of variability, each operated fetus (S or PX) was paired with its unoperated twin (Us or Upx) for plotting and statistical testing. By this means, effects of gestational age differences are removed and effects of genetic differencesare reduced. One-wayANOVAwas also used to test for the presence of significantly different pooled (group) means. Comparisons were made by paired t test between Upx and PX, and between Us and S. When both comparisons showed significant differences, then the effect was assumed to be due simply to the effects of thoracotomy rather than to specific effects of phrenic nerve section. To detect developmental changes over the time course of the experiment (124to 136days gestational age) linear regressions against gestational age were calculated for each variable.
Results
After phrenic nerve section there wereoccasional slow contractions of the diaphragm, but no rhythmic activity and no phasic changes in tracheal pressure. ECoG and nuchal EMG measurements confirmed normal sleep-state cycling. Sham-operated fetuses showed episodic breathing movements in low-voltage ECoG activity, as is normal for fetal lambs of this age. Mean arterial blood gas determinations for group S were: pH 7.406 ± 0.051; Peal' 46.3 ± 6.5; Pal, 18.9 ± 2.02. For the PX group the values were: pH, 7.378 ± 0.070; Pco., 48.1 ± 4.2; POl' 19.0 ± 3.0 (mean ± SD). There were no significant differences between Sand PX. No blood samples were taken from Group U. Mean values ± SD for all measured and calculated variables in the study population are shown in tables 1 and 2. Variables in table 1 showed no significant differences between groups by paired t test or ANOVA, and data from all preparations were pooled. This table includes data from two fetuses not included in the paired testing as they were singleton pregnancies. There was considerable variability between individuals, particularly in DNA measurements and in growth-related variables (e.g., body
1204
BAMFORD, RIVERA, TADALAN, AND ELLIS
TABLE 1 MEAN VALUES OF BIOCHEMICAL AND PHYSICAL MEASUREMENTS IN FETAL LAMB LUNGS* Mean ± SD Fetal wt, g Dry lung wtlbody wt, % V10/kg wet lung wt, ml/kg V10/kg dry lung wt, mllkg V10/kg body wt, ml/kg HPro/wet lung wt, mg/g HPro/dry lung wt, mg/g HPro/DNA, mg/g Elastin/wet lung wt, mg/g Elastin/dry lung wt, mg/g Elastin/DNA, mgimg Elastin/HPro, mgimg Total lung DNA, mg DNA/wet lung wt, mg/g DNA/dry lung wt, mg/g Total lung DNA/body wt, mg/g
n
2,329 ± 615 25 0.42 ± 0.11 25 54±22 26 46±23 26 1.66 ± 0.64 26 1.33 ± 0.28 26 10.15 ± 1.92 26 0.12 ± 0.03 24 1.42 ± 0.42 26 10.92 ± 3.11 26 0.127 ± 0.037 24 1.07 ± 0.26 26 795 ± 243 24 11.11 ± 2.23 24 85.07 ± 17.98 24 0.347 ± 0.089 24
Definition of abbreviations: V10 = volume at 10 em H2 0 ; HPro = hydroxyproline. * Values from all lambs (unoperated, sham, and phrenicnerve-sectioned) were pooled as no significant differences between groups were detected in these variables by ANOVA or by paired t test. Data are included from two fetal lambs (one PX, one S) that were singleton pregnancies and thus were not included in the paired analysis.
weight, compliance), when the standard deviation was 20 to 40070 of the mean value. Hpro determinations showed less individual variability. ANOVA showed significant differences between pooled means for U, PX, and Upx in three variables (absolute VI0, absolute wet lung weight, and dry weight 070). Paired t testing identified the significantly different groups in these comparisons. They were as follows: (1) Absolute wet lung weight was significantly reduced in PX compared with that in Upx (p < 0.02), whereas there was no significant difference between Sand Us. (2) Lung dry weight as a percentage of wet weight showed a small but significant difference between unoperated and operated groups (p < 0.01); they were about 14% higher in PX than in Upxand 12070 higher in S than in Us. However, there was no difference between Sand
PX (figure 1). This effect appears, therefore, to be due to thoracotomy rather than to diaphragm paralysis and lack of fetal breathing. (3) Absolute compliance (VlO) was significantly lower in PX than in Upx (p < 0.02) (figure 2). However, specific compliance, expressed as mllwet lung weight, mlldry lung weight, or mIlkg body weight, showed no significant difference between groups, so the reduced compliance appears to be due to reduced lung size. Us and S groups did not differ significantly. (4) By paired t testing the relative lung weight (wet lung weight/body weight) was significantly lower in PX than in Upx (p < 0.03), whereas there was no difference between Sand Us. Correlation coefficients were also calculated between age at death and each of the variables measured. Surgery was always performed at 110 to 112 days of gestation, so data from all preparations were used in this analysis. When there was no significant difference between groups, group data were pooled together. Over the study period there was a significant correlation (p < 0.05) between pooled values for hydroxyproline per gram dry lung and gestational age, with a slope corresponding to a 35070 increase between 124 and 132 days. Lung dry weight/wet weightdeclined over the study period in all groups, but the correlation coefficient did not reach the 5% level of significance. Elastin and DNA showed no significant change with age, either as absolute values or normalized to dry or wet weight.
20
......
..c:Q.()
'w~
18
...... (l)
~
OWEN S. BAMFORD, AWILDA RIVERA, TERESA TADALAN, and WILLIAM ELLIS Introduction
I f a fetus does not breathe normally in
utero, lung development is delayed, and lungs at birth are smaller and less compliant than normal lungs. Although abnormalities of lung morphometry have been documented in such lungs (1, 2), there is little information on other factors that could contribute to their abnormal mechanical properties. The purpose ofthis study was to examine whether the abnormal mechanical properties are associated with alterations in the deposition of the extracellular structural proteins collagen and elastin. Many workers have examined the effects of experimental abolition or reduction of fetal breathing movements on lung development. Methods used have included bilateral phrenic nerve section (1-3), cordotomy (4), imitation diaphragmatic hernia (5), thoracoplasty to increase thoracic compliance (4), and oligohydramnios (6). Tracheostomy has also been used to bypass the upper airway resistance and therefore to abolish the small distending pressure in the fetal lungs (7). In all these studies, the lungs of the experimental group were smaller and less compliant than those of the control group. Histologic studies of human infant lungs with phrenic nerve agenesis (8), fetal lamb lungs after phrenic nerve section (1), and fetal guinea pig lungs after oligohydramnios (9) all showed severe lung hypoplasia with reduced air space, and similar findings occurred after fetal tracheostomy (7). Thus, there is a consensus that interference with fetal breathing, however caused, affects morphometric and mechanical properties of the lungs. The mechanisms by which fetal breathing movements affect lung properties are less clear. Surfactant does not appear to be involved; surfactant production was not affected by tracheostomy or phrenic nerve section (2, 3, 7). Another possible mechanism is via lung connective tissue. 1202
SUMMARY Interference with fetal breathing movements Is known to retard morphologic development of the lung and to reduce compliance. We hypothesized that the lower compliance might be In part due to effects on lung structural proteins. We studied the effects of phrenic nerve section in utero on lung compliance and on the lung contents of collagen, elastin, and DNA. At 110to 112 days of gestation, one fetal lamb In each of 12 twin pregnancies had either both phrenic nerves cut (PX) or a sham operation (S). The other twin was left unoperated (Upx, Us) as a control. They were killed 14to 22 days later, and the concentrations In lung parenchyma of collagen (as hydroxyproline HPro), elastin, and DNA were measured, together with lung compliance and dry and wet weight. Paired comparisons were made (PX versus Upx and S versus Us). Both operated groups (PX, S) had smaller lungs with lower water content than did their unoperated twins. Absolute static compliance In PX was reduced, but compliance relative to lung weight was unchanged, and there was no significant difference between S and Us. There were no significant effects of PX on the concentrations of HPro, elastin, and DNA, or on the elastin/collagen ratio. Compliance was not correlated with either HPro or elastin content. HPro content Increased significantly with gestational age in all groups. It Is concluded that phrenic nerve section retards the Increase of lung compliance and possibly air space, but It does not affect the overall rate of lung cell proliferation or of deposition of elastin or collagen. Fetal breathing appears to promote physical but not biochemical aspects AM REV RESPIR DIS 1992; 146:1202-1205 of lung development.
The mechanical properties of the lung depend partly on its connective tissue development (10, 11). Schellenberg and coworkers (12) found that both collagen and elastin content of fetal lamb lungs increased during late gestation as compliance increased, and Noguchi and colleagues (13) showed a developmental increase in elastin mRNA in late-gestation bovine lungs. Fetal lamb lungs treated to accelerate lung maturation showed increased lung parenchymal elastin and collagen as well as greater compliance (14). Hypoplastic lungs from human fetuses appear to have less collagen IV in their alveolar septa than do control lungs (15). Thus the increase in compliance might in part be due to increased production of these structural proteins. Conversely, the abolition of fetal breathing could affect lung compliance by reducing their production. In this study we tested the hypotheses that (1) fetuses whose breathing movements are abolished by phrenic nerve section have parenchymal collagen and elastin contents that are reduced compared with those of their unoperated twins, and (2) that the tissue content of one or both of these proteins is cor-
related with lung compliance on air inflation. Methods Surgery Fourteen pregnant Dorset-cross ewes were operated at 110to 112 days of gestation (term, 138 to 140 days), using standard aseptic methods. In the 12twin pregnancies, surgery was performed on only one fetus, the other being left intact as a control. Anesthesia was induced intravenously by 1g sodium pentothal and maintained by ventilation with 20/0 halothane in oxygen. The uterus was approached by a midline incision and opened for access to a fetus, which was withdrawn as far as the umbilicus. The thorax was opened laterally usually between the ninth and tenth ribs, and (Receivedin originalform November 15, 1991and in revised form May 7, 1992) 1 From the Divisionof Neonatology, Department of Pediatrics, University of Maryland, Baltimore, Maryland. 2 Supported by a ResearchGrant from the American Lung Association of Maryland. 3 Correspondence and requests for reprints should be addressed to Dr. O. S. Bamford, Johns Hopkins Children's Center, Eudowood Division of Pediatric Respiratory Sciences, Park 316, 600 N. Wolfe St., Baltimore, MD 21205.
1203
FETAL BREATHING AND WNG STRUCTURAL PROTEINS IN LAMBS
the ribs and lung were retracted separately to expose the inferior vena cava with the phrenic nerve attached. The nerve was lifted and cut at the point where it leaves the vena cava, central to its division over the tendinous portion of the diaphragm. Care was taken to avoid the lymphatic trunks. A pair of stainless steel wire electrodes were sewn into the diaphragm. The ribs and intercostal muscles were then closed by a single layer of sutures, and the skin was closed as a separate layer. The procedure was repeated on the other side. Catheters were placed in the fetal carotid artery and trachea, and recording electrodes wereplaced for electrocortical activity (BCoG) and nuchal extensor electromyogram (BMG). The uterus, body wall, and skin were closed. Sham-operated preparations were treated identically except that the phrenic nerves were identified and lifted but not cut.
Monitoring Fetal arterial pressure and heart rate, ECoG, nuchal and diaphragm EMG, and tracheal pressure were recorded continuously on an eight-channel chart recorder (Grass Instruments, Quincy, MA) until the animals were killed. The diaphragm EMG and tracheal pressure wererecorded to confirm absence of breathing movements. Fetal arterial blood samples were taken daily to check that blood gas and pH values were in the normal range. Measurements At 14 to 22 days postsurgery, preparations were ended by maternal injection of T-61 euthanasia solution (20 ml intravenously). Placental passage of this agent causes deep fetal anesthesia, with loss of ECoG activity within seconds of injection, and cardiac arrest occurs after 1 to 2 min. Compliance. The pressure-volume relationship in static air inflation was used to calculate static lung compliance. Both operated and unoperated fetuses wereimmediately removed from the uterus and weighed, and the thorax was opened carefully to expose the lungs. No attempt was made to remove fluid from the lungs, as it was felt that any such attempt would introduce another uncontrolled variable. A cannula was tied into the trachea, and the lungs were inflated with air from a syringe in 1-mlsteps. To control for the dynamic response of the lung to step-volumechanges, pressure was recorded at a standard time of 30 s after the inflation step, and the next 1ml was immediately added. The process continued until either 10 ml had been added or until the lungs leaked. If there was no leak at 10 ml inflation then air was withdrawn in 1-mlsteps to zero,with pressure measurements continuing as before. Pressure-volume curves were drawn for operated and unoperated lungs, and the volumes at 10 and 20 em H 2 0 were determined graphically (VlO and V20). Only VlO was used for analysis as there was a strong correlation between the two volumes (r = 0.974) and because in one PX and three S fetuses the lungs leaked at 20 em H 2 0 . The phrenic nerves and diaphragm were examined
for evidence of reinnervation. (This was found in only one PX preparation, which showed apparently normal breathing after surgery and was reassigned to the S group.) The lungs were removed, weighed, and frozen for later analysis. Dry weight. Five samples were taken from each pair of lungs (approximately 1 g each sample), one from each right upper, lower, and middle lobe, and one each from the left upper and lower lobe. Samples were weighed wet and then dried at 95° C to constant weight and reweighed to give dry weight. Collagen content. Collagen was estimated from hydroxyproline content of acid hydrolysate. Five tissue samples from the areas of lung used for dry weight wereminced together. Five samples (approximately 100mg each sample) were taken from the mince and hydrolyzed separately in 4 ml6N HCl for 16 h at 100° C. Then 100-lllaliquots of each hydrolysate were dried in vacuum and redissolved in water for hydroxyproline assay by oxidation with chloramine-T at room temperature for 20 min, followed by heating to 60° C for 15 min with Ehrlich reagent. The optical density was then read at 550 nm (16). The assay was calibrated with standard hydroxyproline solutions; the curve was linear over the concentration range found. As a check on the colorimetric hydroxyproline assay, aliquots were taken from 34 hydrolysates from sevenfetuses, dried in vacuum, and redissolved for assay on a Beckman amino-acid analyzer (Beckman Instruments, Fullerton, CA). The correlation coefficient r between determinations by the two methods was 0.94. Elastin content. Insoluble elastin was assayedgravimetrically by the Lansing hot-alkali method, as recommended by Soskel and coworkers (17) for lung tissue. Three 1-gmixed samples were taken from each pair of lungs, minced, and suspended in 30 ml O.lN NaOH in 60 ml polypropylene centrifuge tubes. The tubes were heated in a boiling water bath for 45 min, then cooled and centrifuged for 10 min at 3,000 rpm, and the supernate was poured off. The precipitate was resuspended in cold O.lN NaOH and centrifuged a second time. The supernate was poured off, and the precipitate was washed with cold O.lN NaOH onto an accurately preweighedWhatman GP5 glass fiber disk (Whatman, Clifton, NJ) in a vacuum filtration apparatus. The precipitate was washed twice by filtration with cold O.1N NaOH and then twice more in cold deionized water. Finally, the disk was removed from the filtration manifold, dried in a 60° C oven to constant weight, and weighed on a microbalance accurate to 0.1 mg (Mettler, Hightstown, NJ). DNA content. DNA was measured as an indication of total cell number. Three mixed l-g samples werehomogenized (Cole-Parmer, Chicago, IL) in deionized water to form a 20070 suspension, which was then treated to extract DNA (18). Protein and nucleic acids were precipitated with 10% trichloroacetic acid (TeA) at room temperature; the precipitate was washed with 95% ethanol twice and then
suspended in 5% TeA at 90° C. The protein precipitate was washed twice with 5% TeA, the supernate fractions were combined, and the absorbance was read at 260 nm for DNA concentration.
Statistical Treatment Fetal lambs vary greatly in body weight and developmental stage at both surgery and death. Data on lung development show very large individual variation, and unpaired comparisons between group means can require a prohibitive number of subjects to demonstrate statistical differences. To reduce this source of variability, each operated fetus (S or PX) was paired with its unoperated twin (Us or Upx) for plotting and statistical testing. By this means, effects of gestational age differences are removed and effects of genetic differencesare reduced. One-wayANOVAwas also used to test for the presence of significantly different pooled (group) means. Comparisons were made by paired t test between Upx and PX, and between Us and S. When both comparisons showed significant differences, then the effect was assumed to be due simply to the effects of thoracotomy rather than to specific effects of phrenic nerve section. To detect developmental changes over the time course of the experiment (124to 136days gestational age) linear regressions against gestational age were calculated for each variable.
Results
After phrenic nerve section there wereoccasional slow contractions of the diaphragm, but no rhythmic activity and no phasic changes in tracheal pressure. ECoG and nuchal EMG measurements confirmed normal sleep-state cycling. Sham-operated fetuses showed episodic breathing movements in low-voltage ECoG activity, as is normal for fetal lambs of this age. Mean arterial blood gas determinations for group S were: pH 7.406 ± 0.051; Peal' 46.3 ± 6.5; Pal, 18.9 ± 2.02. For the PX group the values were: pH, 7.378 ± 0.070; Pco., 48.1 ± 4.2; POl' 19.0 ± 3.0 (mean ± SD). There were no significant differences between Sand PX. No blood samples were taken from Group U. Mean values ± SD for all measured and calculated variables in the study population are shown in tables 1 and 2. Variables in table 1 showed no significant differences between groups by paired t test or ANOVA, and data from all preparations were pooled. This table includes data from two fetuses not included in the paired testing as they were singleton pregnancies. There was considerable variability between individuals, particularly in DNA measurements and in growth-related variables (e.g., body
1204
BAMFORD, RIVERA, TADALAN, AND ELLIS
TABLE 1 MEAN VALUES OF BIOCHEMICAL AND PHYSICAL MEASUREMENTS IN FETAL LAMB LUNGS* Mean ± SD Fetal wt, g Dry lung wtlbody wt, % V10/kg wet lung wt, ml/kg V10/kg dry lung wt, mllkg V10/kg body wt, ml/kg HPro/wet lung wt, mg/g HPro/dry lung wt, mg/g HPro/DNA, mg/g Elastin/wet lung wt, mg/g Elastin/dry lung wt, mg/g Elastin/DNA, mgimg Elastin/HPro, mgimg Total lung DNA, mg DNA/wet lung wt, mg/g DNA/dry lung wt, mg/g Total lung DNA/body wt, mg/g
n
2,329 ± 615 25 0.42 ± 0.11 25 54±22 26 46±23 26 1.66 ± 0.64 26 1.33 ± 0.28 26 10.15 ± 1.92 26 0.12 ± 0.03 24 1.42 ± 0.42 26 10.92 ± 3.11 26 0.127 ± 0.037 24 1.07 ± 0.26 26 795 ± 243 24 11.11 ± 2.23 24 85.07 ± 17.98 24 0.347 ± 0.089 24
Definition of abbreviations: V10 = volume at 10 em H2 0 ; HPro = hydroxyproline. * Values from all lambs (unoperated, sham, and phrenicnerve-sectioned) were pooled as no significant differences between groups were detected in these variables by ANOVA or by paired t test. Data are included from two fetal lambs (one PX, one S) that were singleton pregnancies and thus were not included in the paired analysis.
weight, compliance), when the standard deviation was 20 to 40070 of the mean value. Hpro determinations showed less individual variability. ANOVA showed significant differences between pooled means for U, PX, and Upx in three variables (absolute VI0, absolute wet lung weight, and dry weight 070). Paired t testing identified the significantly different groups in these comparisons. They were as follows: (1) Absolute wet lung weight was significantly reduced in PX compared with that in Upx (p < 0.02), whereas there was no significant difference between Sand Us. (2) Lung dry weight as a percentage of wet weight showed a small but significant difference between unoperated and operated groups (p < 0.01); they were about 14% higher in PX than in Upxand 12070 higher in S than in Us. However, there was no difference between Sand
PX (figure 1). This effect appears, therefore, to be due to thoracotomy rather than to diaphragm paralysis and lack of fetal breathing. (3) Absolute compliance (VlO) was significantly lower in PX than in Upx (p < 0.02) (figure 2). However, specific compliance, expressed as mllwet lung weight, mlldry lung weight, or mIlkg body weight, showed no significant difference between groups, so the reduced compliance appears to be due to reduced lung size. Us and S groups did not differ significantly. (4) By paired t testing the relative lung weight (wet lung weight/body weight) was significantly lower in PX than in Upx (p < 0.03), whereas there was no difference between Sand Us. Correlation coefficients were also calculated between age at death and each of the variables measured. Surgery was always performed at 110 to 112 days of gestation, so data from all preparations were used in this analysis. When there was no significant difference between groups, group data were pooled together. Over the study period there was a significant correlation (p < 0.05) between pooled values for hydroxyproline per gram dry lung and gestational age, with a slope corresponding to a 35070 increase between 124 and 132 days. Lung dry weight/wet weightdeclined over the study period in all groups, but the correlation coefficient did not reach the 5% level of significance. Elastin and DNA showed no significant change with age, either as absolute values or normalized to dry or wet weight.
20
......
..c:Q.()
'w~
18
...... (l)
~
