Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The developing fetal lung undergoes dramatic tissue growth and remodeling to achieve the mature alveolar architecture required for optimal gas exchange. The lung remains densely cellular after completing the majority of bronchial branching during the pseudoglandular stage of development. An effective alveolar-capillary interface is then gradually established through the expansion of the endothelial, epithelial, and air space compartments accompanied by a relative thinning of the interstitial compartment. This process begins in utero during canalicular and saccular stages of lung development, and is completed during the largely postnatal alveolar stage. Traditionally, the tissue remodeling has been thought to be due primarily to differences in relative rates of migration and division of cells in the different tissue compartments (1).

It has become increasingly clear that programmed cell death or apoptosis plays an important role in the development of many tissues. The development of the limb into a distal extremity with separated digits (2) and the involution of the tadpole's tail during metamorphosis (3) are prototypical examples. In addition, many internal organ tissues employ apoptosis as a developmental mechanism, including the kidney (4), heart (5), immune (6) and nervous (7, 8) systems. Hormonally responsive adult tissues such as the breast (9), ovary (10), and prostate (11) also utilize apoptosis. Furthermore, the perturbation of normal rates of apoptosis in adult tissues has been associated with a number of malignancies, suggesting an important role for apoptosis in maintaining homeostasis in adult tissues. Apoptosis has been described in the mature lung in response to lung injury (12), but it has not previously been implicated in developing lung.

Apoptosis is characterized by a set of morphologic changes which include detachment of cells from their surroundings, shrinking of the cytoplasm with relative conservation of organelles, condensation of chromatin and fragmentation of the cell and nucleus into well contained fragments called apoptotic bodies. Apoptotic bodies are taken up by macrophages or neighboring cells, presumably allowing the salvage of cell components and avoidance of an inflammatory response. Apoptosis is an active process frequently requiring new protein synthesis (13). Once the process has started, the cell disappears as a morphologic entity in approximately one to two hours. In normal developing tissues that are undergoing substantial levels of apoptosis, the rapidity of the process precludes reliable detection by morphologic methods. Therefore, more specific and sensitive biochemical markers have come into common use as a means of following the apoptotic process. One such marker is DNA digested in a characteristic nucleosomal ladder pattern of multiples of 184 bp of DNA (14). Radionucleotide labeling of the free ends of such DNA increases the sensitivity of this marker enough to allow the detection of apoptosis in normal developing tissues (15). Labeling of the free ends of DNA in situ with detectable tags (i.e., biotin or digoxigenin), greatly increases the sensitivity with which one can localize cells undergoing apoptosis in tissue sections (15). With these and other evolving methods, the role of programmed cell death in normally developing tissues is beginning to be elucidated.

The work reported in this article was undertaken to determine whether apoptosis occurs as a component of the developmental process in rat fetal lung (RFL) and human fetal lung (HFL). We focused primarily on the canalicular and saccular stages during which profound tissue remodeling occurs. In the rat lung, these stages span the last 4 d of gestation and these tissues were readily available. In the human lung, these stages span the 16th to 36th weeks of gestation, but pathologic specimens of human tissue were available for study only during the canalicular stage (15 to 24 wk gestation). An HFL explant culture system had been developed in which both remodeling and epithelial maturation are greatly accelerated (17, 18). The maturation seen in culture is not identical to that seen in vivo (i.e., type I cells remain rare in the face of large scale type II cell development). However, those maturational changes seen in culture that do mimic changes seen in vivo (i.e., growth and maturation of type II cells and interstitial thinning) would take many weeks to occur in vivo. The HFL explant culture system provides a good complement to the in vivo systems because of the accelerated development and because one can evaluate, in vitro, factors suspected to influence the process of apoptosis.

In this article we provide an initial description of apoptosis as it occurs in the rat and human fetal lung in vivo, and in the HFL in explant culture.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Fetal Rat Lung

All studies were approved by the Committee on Animal Research at the University of California, San Francisco. Time-dated pregnant Sprague Dawley rats were obtained from Charles River Labs (Gilroy, CA) and maintained in the UCSF Animal Care Facility for 2-3 d before killing. Pregnant females were anesthetized with a mixture of ketamine (87.5 mg/kg) and xylazine (1.25 mg/kg) intramuscularly, and dams were killed after fetuses were delivered by intracardic injection of 1 ml euthanasia solution containing 67 mg/ml pentobarbitol, followed by bilateral pneumothoraces. Fetuses were delivered at 16 through 22 d gestation; they were pithed and the lungs were dissected. For rats from 18 to 22 d gestation, body and lung weights were obtained. Some lungs were promptly frozen under liquid nitrogen and stored at 70°C for use in biochemical studies. Other lungs were fixed for cryo sections in freshly prepared 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS), pH 7.4. These were fixed overnight, then the lobes were separated and cryoprotected with 30% sucrose in 4% paraformaldehyde/0.1 M PBS, pH 7.4 at 4°C overnight. Individual lobes were placed in O.C.T. compound from Tissue-Tek (Elkhorn, IN), frozen under liquid nitrogen and stored at 70°C.

A third group of lungs was embedded in plastic in preparation for high resolution light and electron microscopy. For these, thin slices of lung tissue were fixed in 2% glutaraldehyde, 1% freshly prepared paraformaldehyde in 0.1 M phosphate buffer (2 h, RT), then postfixed overnight in 1.5% osmium tetroxide in veronal acetate buffer at 4°C. They were stained en bloc in 1.5% uranyl acetate in maleate buffer, then quickly dehydrated in cold acetone and propylene oxide. The tissue was infiltrated and embedded in LX 112 (Ladd Research Industries, Burlington, VT). Semithin (0.5-mm) sections were stained with toluidine blue and examined with a Leitz Orthoplan microscope; thin sections were stained with 5% uranyl acetate and 0.8% lead citrate, then examined in a Zeiss 10 transmission electron microscope.

Adult Rat Brain and Thymus

Three- to four-month old (200-250 g) male Sprague Dawley rats were obtained from Charles River Labs and maintained in the UCSF Animal Care Facility for 2-3 d before killing. They were given intramuscularly 0.6 ml of euthanasia solution which contains 67 mg/ml pentobarbital; this dose has been found to induce deep coma but not to kill the animal. The chest was opened and the thymus removed. The cranium was then opened and a sample of cerebral cortex was removed. Thymus and cerebral cortex were snap frozen in liquid nitrogen and then stored at 70°C until utilized. DNA isolation and labeling for nucleosomal ladder detection were carried out as described below.

Organ Culture of Human Fetal Lung Explants

All studies were approved by the Committee on Human Research at the University of California, San Francisco. Tissue was stored after harvest on wet ice in serum free Waymouth MB 752/1 from GIBCO-BRL (Gaithersburg, MD) and either processed within 2 h of harvest or stored overnight at 4°C and then processed. Lung tissue was minced into 1-mm cubes and cultured in serum free Waymouth MB 752/1 medium in 95% air/5% CO₂, as previously described (17). The tissue was harvested at times indicated, and either snap frozen and stored at 70°C, or fixed and processed for frozen or plastic tissue sections, as described above.

Histology and TUNEL Labeling

The TUNEL (TdT-mediated dUTP-biotin Nick End Labeling of DNA in situ; we use dATP in place of dUTP) method of detecting fragmented DNA in situ was carried out as described (16) with the following modifications. Frozen sections (3-µm thick), cut on a Reichhert-Jung cryostat at 23°C, were treated with proteinase K from Boehringer (Indianapolis, IN) at a concentration of 10 mg/ml for 2 min, rinsed thoroughly, blocked with 1% BSA, 0.03% triton-X100 in 10 mM PBS for 30 min at room temperature (RT) and then rinsed thoroughly again. Labeling with biotinylated ATP from GIBCO-BRL was carried out using the terminal deoxynucleotide transferase (TdT) reaction kit from Boehringer. The reaction mixture contained 200 mM potassium cacodylate, 1.5 mM cobalt chloride, 25 mM Tris-HCl, 0.25 mg/ml BSA (pH 6.6), with TdT at 250 units per ml, and biotinylated ATP at 75 µM. The reaction was carried out at 37°C for 1 h and then stopped with termination buffer (300 mM NaCl, 30 mM NaCitrate). After rinsing, the sections were stained for 30 min with a 1:400 dilution of fluorescein isothiocyanate (FITC)-conjugated streptavidin from Jackson Immunoresearch (West Grove, PA), and then rinsed. The sections were treated with DAPI from Sigma (St. Louis, MO) at a 1:1,000 dilution for 10 min, rinsed, coverslipped with DABCO from Sigma and sealed with nail polish. Slides were photographed with a Leitz Orthoplan epifluorescent microscope through appropriate filter cubes. For cell counting, images were digitally acquired using an Optronics VI-470 CCD camera (Goleta, CA) and Image Pro Plus (Silver Spring, MD) image analysis software. Recorded images were stored on optical disk for subsequent counting, which was done manually. For the apoptotic index (number of apoptotic cells per 1,000 nuclei) of fetal rat lung, we examined two sections of lung from each of two 21-day fetal rats from different dams. The sections were of the entire lung cut in coronal plane. We recorded between 25 and 50 consecutive high power fields (×630) for each section. For each field both DAPI (all nuclei) and TUNEL (only apoptotic nuclei) positive cells were consecutively recorded. We manually counted a total of more than 26,000 cells. A similar procedure was carried out for the HFL explant tissue. We examined 2 sections each from HFL prior to culture and after 1 h of culture. Twenty medium power fields (×400) per section were captured and a total of more than 17,000 cells were counted.

Analysis of DNA

Frozen tissue was homogenized in a Dounce homogenizer in DNA extraction buffer (Tris-HCl 10 mM [pH 7.5], NaCl 60 mM, EDTA 1 mM, proteinase K 0.1 mg/ml, 0.3% SDS) and incubated 16 h at 37°C. The homogenate was extracted first with basic phenol and then with chloroform-isoamyl alcohol. The DNA was then precipitated with ethanol and sodium acetate and resuspended in Tris-EDTA (TE) pH 7.5. The DNA was end-labeled by Klenow polymerase (Boehringer) with ³²P-dCTP from NEN (Wilmington, DE) as previously described (15) with the following modifications: 0.75 µg DNA was treated with 5 units Klenow polymerase using 2 µCi of ³²P-dCTP in 10 mM Tris/ HCl (pH 7.5) and 5 mM MgCl₂ in a total reaction volume of 50 µl. The reaction was incubated for 10 min at room temperature and terminated by adding 25 mM EDTA. The unincorporated nucleotides were removed by 2 consecutive precipitations with ammonium acetate (2.5 M final concentration)/isopropyl alcohol and the DNA was resuspended in 50 µl of TE (pH 7.5). A 20-µl aliquot was subjected to electrophoresis in a 1.8% agarose gel for 3 h at 100 V. After drying the gel between cellophane sheets at 45°C, the gel was exposed for autoradiography using intensifying screens. Aliquots of the reaction mix were also taken for scintillation counting to determine cpm incorporated per µg of DNA.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Rat Fetal Lung

The changes in architecture that occur in the late gestation RFL are illustrated in Figure 1. In the late pseudoglandular stage lung (18 d, Figure 1a), the epithelial tubes have small lumens and there is abundant interstitial tissue. In the late saccular stage lung (22 d, Figure 1b), shortly before natural parturition, there has been an expansion of the epithelial and air space compartments and a thinning of the interstitium; although, compared with fully mature lungs, the interstitium was still relatively cellular (not shown). During this period the RFLs showed considerable growth: the average wet weight increased from 24 ± 5 mg (mean and SD) at 18 d to 127 ± 20 mg at 22 d gestation. Figure 2 is an electron micrograph of RFL at 18 d. Densely stained cell fragments with typical features of apoptotic bodies (arrows) are shown in the interstitium. Interestingly, one also sees a mitotic interstitial cell in close proximity to the apoptotic bodies. Formal cell counts were not done at the electron micrograph level; apoptotic bodies were seen only rarely and then only in the interstitium.

View larger version (71K): [in this window] [in a new window]   Figure 1.   Light micrograph showing development of rat fetal lung in vivo. (Panel a) Rat fetal lung at 18 d shows small lumens and abundant interstitial tissue. (Panel b) Rat fetal lung at 22 d gestation shows greatly increased epithelial and air space compartments and relatively thinner interstitial tissue. Bar = 100 µm.

View larger version (192K): [in this window] [in a new window]   Figure 2.   Electron micrograph of rat fetal lung at 18 d gestation. Arrows indicate cells or cell fragments that contain very dense chromatin and show morphologic features consistent with cells undergoing apoptosis. The asterisk marks a nearby cell with condensed chromosomes and dissolved nuclear membrane in the process of mitosis. Bar = 5 µm.

The nucleosomal ladders pattern of DNA degradation that is a hallmark of apoptosis is seldom detectable by ethidium bromide staining under physiologic conditions, therefore we adapted a very sensitive method based on radioactive end labeling of extracted DNA. To assess the range of nucleosomal ladder signals that could be detected with the method of DNA analysis which we employed, we compared equal masses DNA from 16-wk fetal lung, young adult rat cerebral cortex, and adult rat thymus; the results are shown in Figure 3a. The fetal lung shows a strong ladder pattern. The brain (B) shows no convincing ladder pattern, but minimal signal in the broad 200 base pair region. Linear densitometry scanning of the last 5 rungs of the DNA ladder in the RFL lane and the comparable region in the adult rat brain lane showed a ratio of 400:1. The thymus showed a very strong signal (lane TL) that was best seen at a shorter exposure (lane TS).

View larger version (55K): [in this window] [in a new window]   Figure 3.   (Panel a) Autoradiograph of 32P-labeled DNA from 16 d gestation fetal rat lungs (L), from young adult rat brain (B) and thymus (TL in the longer exposure of 24 h, TS in the shorter exposure of 8 h). Fetal lung and adult thymus showed clear nucleosomal ladders. A standard 123 bp DNA ladder was also labeled (S) and run alongside the experimental samples. (Panel b) Autoradiograph of 32P-labeled DNA from seven fetal rat lungs harvested at (a) 16, (b) 17, (c) 18, (d) 19, (e) 20, (f) 21 and (g) 22 d gestation. All samples showed a nucleosomal ladder. The positions of the 123, 246, 369 and 738 bp bands of the 123 bp DNA molecular weight marker are shown at the left of the figure.

We employed this method to show DNA digested in a nucleosomal ladder pattern characteristic of apoptosis in RFL of gestational ages 16 through 22 d (Figure 3b). DNA from different lungs showed some variation in the intensity of the nucleosomal ladder but, at the resolution of daily sampling, there was not a consistent relationship to gestational age. Variation was seen between DNA of littermates, but not between samples of DNA from a single lung when multiple aliquots were labeled and fractionated separately (data not shown). The 16 d RFL DNA shown in Figures 3a (sample [L]) and 3b (sample [a]) were equal aliquots of the same DNA extract, labeled in the same way but at different times, as the DNAs from each figure were labeled together; the exposure time in Figure 3a was longer than 3b (24 versus 18 h).

Labeling of the digested DNA in situ by TUNEL staining of 21 d RFL, demonstrated positive cells that appeared to be located in the interstitial tissue compartment. The apoptotic indices in two different 21 d RFLs (two sections from each) from different dams were 1.9 (range 1.5 to 2.1), and 2.6 (range 2.4 to 2.7) positive cells per 1,000 nuclei. Labeled nuclei were seen in a variety of forms including whole evenly stained nuclei, nuclei showing perinuclear condensation of labeled DNA, and labeling of small cell fragments known as apoptotic bodies (Figure 4).

View larger version (118K): [in this window] [in a new window]   Figure 4.   Section of rat fetal lung at 21 d gestation stained by the TUNEL reaction to detect apoptotic cells (green fluorescent nuclei in panels a and d). The section was also stained with DAPI to identify all nuclei present (blue fluorescent nuclei in panels b and e). Phase contrast views of the tissue are also shown for orientation (panels c and f  ). The top panels (a, b and c) are lower magnification views which show the typical distribution of apoptotic cells in the tissue. Bar = 100 µm. The positive nuclei in panel a are all interstitial and range in size from whole nuclei to nuclear fragments (apoptotic bodies). The lower panels (d, e and f ) are higher magnification views which illustrate morphologic detail. Bar = 40 µm. The positive nucleus in panel d shows the marginated fragmented DNA typical of apoptosis. Red fluorescence represents the autofluorescence of endogenous cells, most noticeably red blood cells.

Human Fetal Lung Preculture

DNA digested in a nucleosomal ladder pattern characteristic of apoptosis is also present in HFLs of 15 through 24 wk gestation (Figure 5). The overall intensity of the DNA ladders, as well as the variation in the intensity of the ladders, were of the same general magnitude for the HFLs as it was for the RFLs. There was no consistent relationship between ladder intensity and gestational age at the resolution of weekly sampling. Ladder intensity was not influenced by time of tissue storage before culture between 2 and 24 h after harvest (data not shown). TUNEL data are discussed below.

View larger version (115K): [in this window] [in a new window]   Figure 5.   Autoradiograph of 32P-labeled DNA from eight human fetal lungs obtained at (a) 15, (b) 17, (c) 19, (d) 20, (e) 21, (f) 21, (g) 24 and (h) 24 wk gestation. DNA was extracted, labeled with 32P, and fractionated on a 1.8% agarose gel. A nucleosomal ladder pattern of DNA degradation is clearly apparent. The positions of the 123, 246, 369 and 738 bp bands of the 123 bp DNA molecular weight marker are shown at the right of the figure.

Human Fetal Lung in Explant Culture

A comparison of 20 wk HFL on day 0 and day 4 of explant culture (Figures 6a and 6b, respectively) showed an expansion of potential air spaces and epithelial cell surface area with a concomitant decrease in the relative amount of interstitial tissue present. These changes superficially resemble the changes seen in the developing RFL (compare Figure 1). On closer examination, one sees important differences. The most pertinent of these differences was the surprising number of apoptotic interstitial cells in the HFL (Figure 7). Also notable was an epithelium in the HFL that consists almost exclusively of cuboidal cells (Figures 6b and 7), which at the EM level are seen to have microvilli and lamellar bodies (Figures 7a, 7b) (17, 18), both features of the type II pulmonary epithelial cell.

View larger version (76K): [in this window] [in a new window]   Figure 6.   Morphologic changes seen in human fetal lung explants after 4 d in culture. (Panel a) Preculture specimen shows small lumens and abundant interstitial tissue. Epithelium is immature with abundant glycogen. (Panel b) Tissue after 4 d in culture shows greatly increased epithelial and air space compartments, relatively thinner interstitial tissue and a uniform cuboidal epithelium, with reduced glycogen, lining the air spaces. Bar = 100 µm.

View larger version (130K): [in this window] [in a new window]   Figure 7.   Human fetal lung after 4 d in explant culture showing evidence of apoptosis. (Panel a) Light micrograph. Present in the interstitial tissue, but not epithelium, are cells undergoing apoptosis characterized by perinuclear chromatin condensation (arrows). Bar = 20 µm. (Panel b) Electron micrograph. Note the apoptotic cell characterized by condensed chromatin and shrunken cytoplasm in the lung interstitium (arrow) surrounded by intact, developing lung tissue. Epithelium shows osmiophilic bodies and microvilli. Bar = 10 µm.

When HFL explants were placed in culture, they showed a very rapid increase in the number of cells undergoing apoptosis. This was seen most clearly by analysis of the DNA extracted from tissue sampled as a function of time in culture. DNA digestion was first evident in the high molecular weight (MW) fraction which showed increased labeling within 1 h. Labeling then progressively increased and shifted to the lower MW bands of the DNA ladder. Concomitantly the high MW DNA showed a decreasing signal. This pattern is illustrated in Figure 8 for 3 lungs of 15, 19 and 21 wk gestation that were harvested, cultured and processed at the same time. A similar pattern has been seen in each of the 10 lungs, gestational ages 15 to 24 wk, that we have studied.

View larger version (104K): [in this window] [in a new window]   Figure 8.   Autoradiograph of 32P-labeled DNA from three human fetal lungs of gestational ages 15, 19, and 21 wk. The lungs were harvested after 0, 1, 4 or 18 h in explant culture. DNA was extracted, labeled with 32P, and fractionated on a 1.8% agarose gel. Samples from the three gestational ages are grouped together for comparison at each time of harvest. Note the differences in the time of onset and rate of progression of the DNA ladder for the three gestational ages. At 0 h, at this exposure, only the 21 wk lung shows labeled large MW DNA. At 18 h large MW DNA from the 19 and 21 wk lungs is no longer detected by labeling, but it is still seen in the 15 wk lung.

Although the overall pattern of DNA digestion that occurs during the wave of apoptosis induced by explant culture is invariant, the rate of progression is not. In Figure 8, one sees that in the preculture tissue (time 0) an equal mass of DNA from the 15 and 19 wk lungs have less intense end labeling, less apoptosis, than the 21 wk fetal lung (all three can be seen to be labeled on longer exposure). Furthermore, the rate of progression of the nucleosomal DNA ladder into small molecular weight rungs seems to be faster in the later gestation lungs, and somewhat independent of the intensity of preculture labeling (compare 19 and 21 wk lungs with the 15 wk lung). We replicated this relationship between the rate of progression of DNA digestion and gestational age in one other experiment with 18 and 24 wk HFLs that were harvested, and had their DNA analyzed, at the same time (data not shown). We have also determined that the rate of DNA digestion did not vary with storage times ranging from 2 to 24 h after harvest (data not shown). Figure 9a shows ethidium bromide staining of extracted DNA from tissue cultured for 3 d. Ethidium bromide detects DNA mass instead of the number of free DNA ends; one misses the early digestion of DNA, but can observe an increasing proportion of DNA in the low molecular weight rungs of the ladder (from apoptotic cells) for the first day or two in culture followed by an unchanging distribution of DNA for the remainder of the experiment. Intact DNA (from living cells) in the high MW range persisted for the duration of the culture even though end labeling in this region stops by 24 to 48 h.

View larger version (121K): [in this window] [in a new window]   Figure 9.   (Panel a) Ethidium bromide-stained DNA from a cultured human fetal lung of gestational age of 20 wk sampled after (a) 0, (b) 1, (c) 2, (d) 4, (e) 6, (f) 10, (g) 18, (h) 22, (i) 25, (j) 48 and (k) 72 h of culture. DNA digestion by this method of detection is just detectable at 6 h of culture, seems to be complete by 48 h of culture and then remains unchanged out to 72 h of culture. Note large molecular weight DNA in the exclusion area of the gel is present out to 72 h of culture. (Panel b) Black and white image of TUNEL staining of human fetal lung explant tissue cultured as described for 4 d. Note a number of positive cells in the interstitium of the tissue.

The cells in the HFL explants that contained the fragmented DNA were localized in situ using the TUNEL method (Figure 10). As expected, in the preculture tissue one sees rare labeling of cells and these appear to be restricted to the interstitium (Figure 10a). The apoptotic index of one preculture HFL (two sections) was 2 (range 1.8 to 2.2) per 1,000 nuclei, very close to that for developing RFL at 21 days gestation. After only 1 h in culture, TUNEL staining revealed widespread labeling of well circumscribed cell nuclei (Figure 10d), paralleling in situ the increased labeling seen in the extracted DNA (compare Figure 8). The apoptotic cells appeared to be confined to the interstitium and the apoptotic index from one 1 h of culture HFL (two sections) was 110 (range 99 to 120 per 1,000 nuclei), a 55-fold increase compared with time 0. Interestingly, labeled cells can still be seen in large numbers in the interstitium on day 4 of culture (Figure 9b). At both 1 h and 4 d of culture, most labeling was seen in whole nuclei. In contrast, on TUNEL staining of the in vivo RFL and preculture HFL, labeled fragments, much smaller than whole nuclei, were more often seen.

View larger version (118K): [in this window] [in a new window]   Figure 10.   Human fetal lung explant tissue sampled at times 0 and 1 h of culture. The sections were stained by the TUNEL reaction to detect apoptotic cells (green fluorescent nuclei in panels a and d) and with DAPI to identify all nuclei present (blue fluorescent nuclei in panels b and e). Phase contrast views of the tissue are also shown for orientation (panels c and f  ). Time 0 shown in top panels (a, b and c). Few fluorescein positive interstitial cells are present in panel a while panel b shows a dense field of DAPI positive cells. Arrow marks a positive cell in panels a, b and c. Time 1 h shown in bottom panels (d, e and f  ). Panel d documents the large increase in TUNEL positive cells after the first hour of culture. Arrows mark three examples of positive cells in panels d, e and f. Apoptotic cells are seen to be in the interstitium in panels a and d. Panels e and f show morphology similar to panels b and c, respectively. Red fluorescence represents the autofluorescence of endogenous cells, most noticeably red blood cells. Bar = 100 µm.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We hypothesized that apoptosis was one of the developmental forces that shape the developing lung. We demonstrated the morphologic changes that occurred in the RFL between 18 d and 22 d gestation and in the mid trimester HFL over 4 d of explant culture and showed that they underwent superficially similar levels of epithelial and air space expansion and interstitial condensation. We then employed three well established methods used to study apoptosis to develop consistent and converging evidence that there were cells in these developing fetal lungs that were undergoing apoptosis.

Electron microscopy provided the most informative and definitive means of detecting apoptosis, though it was not suitable for quantitative analysis of rare apoptotic cells in tissue. An electron microscopy image from 18 d RFL was unequivocal and the occurrence of cell replication and cell death, in this single electron microscopic field, illustrated well the fundamental point that these seemingly opposite cell choices often occur in proximity as part of the dynamic process of development (19). We did not extensively search uncultured HFL by EM for rare apoptotic bodies but did demonstrate abundant interstitial cells with apoptotic morphology by light and electron microscopy in cultured HFL. We observed in the same microscopic fields other, healthy appearing interstitial cells and healthy appearing and differentiating epithelial cells.

The nucleosomal ladder pattern seen on agarose gel electrophoresis of DNA either stained with ethidium bromide for mass or radioactively end labeled to detect the number of free ends, proved to be a versatile way of verifying the occurrence of apoptosis. End labeling is known to be 1,000 to 2,000 times more sensitive than ethidium bromide staining (15). With end labeling, we have seen nucleosomal ladders in tissue with less than 1 apoptotic cell per 1,000. We analyzed DNA from thymic and cerebral cortical tissue from young adult rats (about 3 months) as relative positive and negative controls, respectively. As expected, the thymic tissue showed an intense nucleosomal ladder pattern. We could not demonstrate a convincing ladder pattern for the brain tissue but there was some signal in the low MW region that was a minimum of 400-fold less intense than that seen in a RFL sample. The reported apoptotic index for young adult rat brain is 2 per 100,000 cells (20), which is about 100-fold less than we report for RFL.

At the onset of apoptosis, large MW DNA shows radioactive end labeling (DNA digestion) first, with subsequent progression to the characteristic nucleosomal ladder pattern (21). Signal in the lower MW bands is frequently used for quantification (i.e., 22). This can be problematic, especially when evaluating apoptosis in tissue, because the intensity of labeling in those bands is dependent not only on the number of cells undergoing apoptosis, but also on the rate of DNA digestion within those cells and the time the digestion proceeds before the cell fragments are consumed and fully broken down.

The distribution of DNA in the ladders from cultured HFL tissue was different from that seen in vivo. Within 1 h of being placed in culture, HFL showed a sharp increase in the end labeling of high MW DNA (early apoptosis), which was then followed by a progressive increase in the low MW rungs of the ladder. The labeling of high MW DNA returned to baseline by 24 h, even though there was still abundant high MW DNA present, as seen by ethidium bromide staining. This suggested that the increase in apoptosis occurred over several hours and then decreased to baseline. Moreover, in the absence of new high MW DNA end labeling, the low MW DNA remained present for several days. This observation and the abundant apoptotic interstitial cells detected on Day 4 of culture by TUNEL labeling and electron microscopy lead us to hypothesize that the clearance of apoptotic cells was impaired in the HFL explant culture system.

In contrast, in the preculture HFL and in vivo RFL, the intensity of the nucleosomal ladder suggested that there was a lower, seemingly more steady, level of apoptosis occurring. There was a relative balance in the end labeling of high and low MW DNA, suggesting, perhaps, a balance between onset of apoptosis and clearance of apoptotic bodies. Occasionally in the fetal lung in vivo, intense nucleosomal DNA ladders were seen, possibly representing periodic increases in the level of apoptosis that were out of synch with our schedule of lung sampling. It is clear that more detailed and quantitative analyses of the pattern of apoptosis are required to resolve these issues.

The TUNEL method provided a means of identifying apoptotic cells in histologic sections and thereby permitted the determination of the number of cells undergoing apoptosis and the localization of those cells to a tissue compartment (interstitium versus epithelium). The sensitivity of TUNEL is limited only by the number of fields one is willing to scan. The control samples, which lacked only the labeling enzyme, were negative for fluorescently labeled cells. Therefore, we are confident that labeled cells contained fragmented DNA, a finding supported by the presence of the nucleosomal ladder pattern of DNA degradation in the electrophoresis data. Moreover, there was broad agreement between the DNA gel electrophoresis and TUNEL analyses that is most apparent in the HFL explant culture system where the changes in the level of apoptosis were consistent and large. Although quantitative morphologic techniques were not used in the determination of the apoptotic indices, we examined a large number of consecutive fields which included both peripheral and proximal lung tissue. Therefore, we believe the values we obtained for the apoptotic indices for RFL and preculture HFL are representative. They are similar to each other and in the range of those seen in other tissues in which apoptosis is effecting important changes (i.e., the developing nervous system [23]). If one assumes a one hour detection span for apoptotic cells (i.e., 24) and a constant rate of clearance, then an apoptotic index of 2 apoptotic cells per thousand over 4 d would result in a loss of about 19 percent of all cells. Therefore, it seems likely that apoptosis is a significant factor in the remodeling of the developing lung. Periodic increases in the rate of apoptosis missed due to limited spatial and temporal sampling was suggested by the DNA ladder data. This would imply a further impact by the apoptotic component of lung remodeling.

In the electron microscopic fields that we examined, all apoptotic cells were definitely in the interstitium, as exemplified in Figures 2 and 7b. This was also true for high-powered light microscopy of 0.5-micron thick plastic embedded tissue sections, as exemplified in Figure 7a. In a survey of many fields in 19 and 21 d RFL epithelial cells in these plastic sections, no pyknotic nuclei were detected, whereas they were easily detected after certain treatments which induce epithelial cell death (data not shown). In the fixed frozen sections examined with light microscopy, the majority of apoptotic cells appeared convincingly to be confined to the interstitium. However, the resolution under these circumstances is not adequate for certainty in all cases. Furthermore, because in the rat and human only about 20% of lung cells are in the epithelium (25), a low rate of apoptosis in the epithelium could have been missed. Much less than 18% of the TUNEL positive cells were equivocal as to their location in the interstitium, suggesting the epithelium is spared in the developmental contexts that we evaluated.

Although this report marks the first clear documentation of the occurrence of apoptosis in developing lung in vivo and in vitro, it just scratches the surface of the work that needs to be done to fully define the importance of apoptosis for lung development. Among the many questions still to be defined are the magnitude of the impact on the lung, the period of development during which it is most important, and the effects on the process of perturbations in normal development that are so much a part of modern perinatal-neonatal medicine. Also, the factors which control the process in vivo and in vitro need to be defined, as do the effector pathways employed by the developing lung cells. Finally, the role of disrupted apoptosis in congenital malformations of the lung needs to be explored.