Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Lung growth appears regulated largely by factors that influence pulmonary distension (1). A reduction in thoracic volume either experimentally induced or naturally occurring retards lung growth in both animals and humans (1, 2, 5). There is good evidence both in rabbits and sheep that the mechanical forces caused by fetal breathing movements have a profound influence on lung growth (2, 6, 7). Many published case reports indicate that full-term newborns with in utero akinesia, such as severe congenital myopathy, have pulmonary hypoplasia that can lead to respiratory complications with morbidity and mortality (8, 9). Evidence from autopsy studies suggests that in utero breathing movements greatly influence human lung growth (5). Mechanical strain has also been shown to upregulate DNA synthesis (10, 11) as well as differentiation (12, 13) in cultured lung cells. Although the importance of fetal breathing movements on lung growth seems relatively clear, there remains uncertainty as to the influence of fetal breathing movements on maturation of lung cells in vivo (2, 14).

The secondary effects of surgical interventions to alter fetal breathing movements (5, 18) have confounded in utero evidence for a role of mechanical distention on lung development. Such experimental perturbations, e.g., cervical transections, are fraught with a high mortality rate (> 50%) and result in some oligohydramnios that alone can cause fetal pulmonary hypoplasia (13). Furthermore, cervical transection cannot completely eliminate electromyography activity of involved skeletal muscles as seen with phrenic nerve denervation (21). Also transection likely leads to increased catecholamine and cortisol concentrations that can influence pneumocyte type II differentiation (22, 23). Usage of a noninvasive model to diminish fetal breathing movements in a proportion of embryos in a given litter in utero (2, 5, 18) would eliminate some of these confounding variables and would likely provide a different perspective.

Myogenin (24, 25) belongs to the MyoD family of basic helix-loop-helix (bHLH) transcription factors that activate the skeletal muscle differentiation lineage (26). As a transgenic mouse model of a severe congenital myopathy, the myogenin knockout (homozygous null /) mice lack normal skeletal muscle fiber formation (27). An examination of their lungs after birth showed that they had never inspired air (27, 28), indicating no breathing movements. In contrast, littermates (heterozygous +/ and homozygous +/+) develop with normal skeletal muscle and breathing movements (27).

In the mouse, lung development has been divided into four stages: pseudoglandular (embryonic days [E]9.5- E16.6), canalicular (E16.6-E17.4), terminal sac (E17.4- postnatal Day 5), and alveolar phase starts about Day 5 (29). E20 is considered full term. For the diaphragm, E14 heralds normal formation of skeletal muscle fibers and marks the earliest time contractile activity may commence (30, 31). Based on all of these facts, we hypothesized that the absence of skeletal muscle and fetal breathing movement in utero from E14 to E20 in myogenin null mice would result in lung hypoplasia, decreased lung cell proliferation, and increased apoptosis in the lungs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animal Care

All mice were housed in an animal facility at 22°C with a 12:12 h light:dark photoperiod. The Institutional Animal Welfare Committee, University of Texas Health Science Center at Houston, approved all our animal protocols. The phenotype of myogenin null (/) mice has been described previously (27). For breeding, heterozygous males and females were intercrossed. The presence of vaginal plugs indicated the beginning of gestation (E0). On E12, E14, E17, and E20 (full term), pregnant females were killed by cervical dislocation, and embryos/fetuses were removed by cesarean section. Each fetus was weighed and processed for further analysis.

Unlike the perinatal lethal phenotype of myogenin homozygous null pups, wild-type and heterozygous littermates are viable and phenotypically normal. Withdrawal reflexes could not be elicited in any myogenin null E17, E19, or term pups.

Embryonic organs (e.g., lung, heart, liver, limbs) were dissected for RNA extraction. Skin, lung, or tail tissues were taken for DNA genotyping by Southern blot analysis as described previously (27). Although we cannot exclude the possibility that the homologous recombination "knockout" affects flanking gene(s) (32) that are involved more directly with lung development, this effect seems unlikely.

Lung Weights and Total DNA

Lungs of fetuses were blotted, weighed, lyophilized > 48 h and reweighed. The dried lungs were homogenized in 10% trichloroacetic acid followed by perchloric acid then chloroform extraction for DNA determination (33). Tissue content of total DNA was based on whole lung and dry weights. Genomic DNA was also extracted with a modified method using a 10% Triton X-100 DNA assay with Hoechst in a DNA spectrophotometer (GeneQuant, Amersham Pharmacia Biotech, Piscataway, NJ) (34).

Tissue Preparation

Tissues were prepared for analyses by various methods depending on previously published specific conditions. Wax-embedded tissues were used for immunolabeling with 5'-bromo-2'-deoxyuridine (BrdU) and Bax antibodies (35, 36). Tissues flash frozen with liquid nitrogen were employed for surfactant protein (SP)-A immunofluorescence (our unpublished observation). Sucrose-embedded frozen tissues were made for the F5D antibody (37). Freshly dissected tissues were placed in Carnoy's fixative for 28 to 30 h for routine paraffin embedding. Sections 5 µm thick were cut with a microtome and placed on chromium-potassium gelatin-coated slides. Sections were air-dried overnight on a 37°C slide dryer. Routine hematoxylin and eosin (H&E) staining of serial sections enabled determination of intrathoracic space and lung volume with histologic reconstruction as previously reported (33, 38).

Lung Proliferation In Vivo

A total of 10 mg of BrdU from Sigma (St. Louis, MO) was freshly dissolved in 1 ml sterile normal saline. A bolus of this BrdU solution was injected (10 µl/g body weight) intraperitoneally in pregnant females 2 h before sacrifice (35). A segment of the mother's intestine was excised with each litter for positive BrdU control tissue. Embryo heart tissue adjacent to lung tissue served as internal controls to show proliferation activity, particularly in late gestation and newborn stages.

Antigen Retrieval for BrdU, Bax, Proliferating Cell Nuclear Antigen, and TUNEL Staining

To enhance antigenicity, tissue sections were deparaffinized, boiled in 0.1 M citric acid buffer, pH 6.0, for 5 min (39), and neutralized with 3% H₂O₂ for 15 min at room temperature to inhibit endogenous peroxidase activity.

Immunolabeling of Tissue BrdU and PCNA

The sections were rinsed three times in phosphate-buffered saline (PBS) and then blocked in 0.1% nonfat dry milk in PBS for 30 min. Sections were then incubated overnight at 4°C with 1:50 mouse anti-BrdU (G3G4) monoclonal antibody purchased from the Developmental Studies Hybridoma Library (Iowa City, Urbana, IA; originally provided by Dr. Steve J. Kaufman, Illinois) (35). After three washes in PBS (5 min each), appropriate secondary antibody (Zymed, San Francisco, CA) conjugated with fluorescein isothiocyanate or horse-radish peroxidase was applied for 4 h at room temperature. Negative control samples (omitting primary antibody) were run in parallel. Sections were also immunolabeled with anti-proliferating cell nuclear antigen (PCNA) (Zymed) and results were found to be comparable to BrdU incorporation (data not shown) in the lung as well as maternal intestinal crypt cells. In random fields, all BrdU positive cells in 12 photomicrographs (×400 magnification) were counted relative to non-BrdU, Hoechst-DNA stained cells.

Bax Immunoperoxidase

Rabbit antihuman/mouse Bax polyclonal (Santa Cruz Biotech, Santa Cruz, CA) was diluted 1:100 and stained as previously described (36).

TUNEL

Sections were incubated for 1 h at 37°C with terminal deoxynucleotide transferase (TdT) and biotin-labeled deoxyuridine triphosphate (dUTP) (35, 40). After incubation, incorporated nucleotide was detected using avidin-peroxidase coupled antibody with diaminobenzidine (Sigma) as substrate. Sections were counterstained with hematoxylin for nuclear DNA. As a control for nonspecific enzyme activity, some samples were incubated without TdT. Serial sections were pretreated with 0.01% DNAse as positive controls. To quantitate results, all of the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL)-positive and TUNEL-negative lung cells were counted and scored in 12 mosaic micrographs of a section of a single null and wild-type littermate as described previously for BrdU labeling.

SP-A Immunoperoxidase Labeling

Rabbit antihuman SP-A polyclonal antibody (gift of Dr. Carole R. Mendelson, Department of Biochemistry, University of Texas-Southwestern, Dallas, TX) was diluted 1:50. Three washes of PBS (5 min each) were applied to air-dried, flash-frozen tissue sections and then the appropriate secondary (Zymed) antibody, which was conjugated with horse-radish peroxidase, was applied for 4 h at room temperature. Negative control samples counterstained for nuclear DNA (omitting primary antibody or employing preimmune serum) were run in parallel.

Myogenin Immunoperoxidase Labeling with F5D

For E12 and E14 specimens, flash-frozen, sucrose-embedded tissues were prepared as described previously (37). Sections were stained with mouse antirat myogenin (F5D) monoclonal antibody (25) (diluted 1:100) that had been purchased from the Developmental Studies Hybridoma Library (originally provided by Dr. Woodring Wright, University of Texas Southwestern Medical Center, Dallas, TX).

Electron Microscopy

Freshly dissected lung tissues (< 2 mm³) were fixed in 3% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4; 350 mOsm), followed with 1% osmium tetroxide. Blocks were dehydrated, embedded in Epon 812, cut to 80 nm, placed on Parloidin-coated copper grids, stained with uranyl acetate and Reynold's lead citrate, and observed on a JEOL JEM-100EX (JEOL USA, Inc. Peabody, MA).

RNA Isolation and RT-PCR

Total RNA was isolated by the guanidine isothiocyanate method (Trisolve reagent). A total of 1 µg of total RNA was reverse transcribed with random hexamer primers, then reverse transcriptase/polymerase chain reaction (RT-PCR) amplification of complementary DNA (cDNA) was performed with the described oligonucleotides that were designed to cross introns so that contaminant DNA could be detected as a larger PCR product. The following primers were employed: myogenin forward primer, 5'-ACC AGG AGC CCC ACT TCT AT; myogenin reverse primer, 5'-CAT CAG GAC AGC CCC ACT TA (30 cycles [95°C for 1 min, 58°C for 1 min, and 72°C for 2 min] melting temperature [Tm] = 64°C; 722 bp RNA, 1763 bp DNA); SP-A forward primer, 5'-ATC CCT GGA GCT CCT GGA AA; SP-A reverse primer, 5'-ACA GAA GCC CCA TCC AGG TA (30 cycles [95°C for 1 min, 55°C for 1 min, and 72°C for 1.5 min] Tm = 65°C; 552 bp RNA); SP-C forward primer, 5'-ATT ACT CGG CAG GTC CCA GGA GCC A; SP-C reverse primer, 5'-AGA TAT AGT AGA GTG GTA GCT CTC C (30 cycles [95°C for 1 min, 55°C for 1 min, and 72°C for 1.5 min] Tm = 66°C; 574 bp RNA); S16 forward primer, 5'-AGG AGC GAT TTG CTG GTG TGG A; S16 reverse primer, 5'-GCT ACC AGG CCT TTG AGA TGG A (26 cycles [94°C for 1 min, 60°C for 1 min, and 72°C for 1 min] Tm = 60°C; 125 bp RNA). The primer sequences for SP-A and SP-C are taken from Cardoso and colleagues (41, 42) and the S16 primers were taken from Leonard and coworkers (43). All oligonucleotide primers were synthesized by Genosys Biotechnologies (Houston, TX).

Statistics

Paired groups of data were analyzed by Student's t test using the Systat software (SPSS, Inc., Chicago, IL) (44). Results were determined to be significant if P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

myogenin Null Pups Become Cyanotic and Die Shortly after Birth

The myogenin null pups, unlike their wild-type littermates, were born without any spontaneous limb or respiratory movements, although their hearts were beating. Initially, myogenin null pups had pink color similar to their littermates; however, myogenin null pups, lacking the ability to ventilate, became cyanotic within seconds of birth (Figure 1) and died within minutes. Unlike wild-type littermates, myogenin null pups exhibited a profound kyphosis (Figure 1), likely owing to the lack of sufficient vertebral skeletal musculature during development.

View larger version (122K): [in this window] [in a new window]   Figure 1.   Heterozygous littermate (+/) and myogenin null (/) pups shortly after birth. At birth, myogenin null pups were the same color as heterozygous littermates, but within seconds of birth, they became cyanotic.

myogenin Null Pups Have Lung Hypoplasia

Three criteria indicate that myogenin null pups have lung hypoplasia. First, the ratio of wet lung weight to body weight was lower in the myogenin null mice at E14, E17, and E20 (Table 1). The definition of pulmonary hypoplasia has been previously reported to be wet lung weight to body weight that is < 67% of normal at the time of birth (33). It is clear from these time points that the myogenin null embryos incur pulmonary hypoplasia throughout the E14 to E20 time frame. Nevertheless, overall organogenesis of the mouse lungs is maintained; four right lung lobes and a single left lung lobe flanking the heart are easily seen (Figure 2A). Second, complete sagittal serial reconstructions of whole E14 embryos stained with H&E to assess intrathoracic volumes revealed the null embryos had approximately 30 ± 11% (standard deviation [SD]) smaller intrathoracic spaces as compared with age-matched littermates (n = 6; data not shown). Third, histologically the myogenin null lungs at term appeared to be arrested in a canalicular stage where respiratory epithelium was cuboidal (Figures 2B and 2C); however, the wild-type littermates had achieved the terminal sac stage where septal walls continued to be thin, and cuboidal epithelia became thin squamous types.

View larger version (88K): [in this window] [in a new window]   View larger version (122K): [in this window] [in a new window]   Figure 2.   (A) Lungs and heart (H) of heterozygous littermates (+/) and myogenin null (/) pups at E14 (upper panel) and E20 (lower panel). A millimeter scale is shown above each age. Wax-embedded lung sections stained with H&E from heterozygous (+/) littermates (B) and myogenin null (/) (C) pups. At term (E20), myogenin null lungs appeared to be arrested in a canalicular stage where the epithelium has become cuboidal as compared with heterozygous or wild-type littermates, in which septal walls have thinned and terminal sacs have formed. Distance of bar is 50 µm.

Lung Hypoplasia Is Not Due to the Absence of Myogenin Expression in Lungs during Development

In wild-type lungs, myogenin messenger RNA (mRNA) and protein were not detected by RT-PCR and immunolabeling analyses, respectively, at E12, E14, or E17 (Figures 3A and 3B), whereas myogenin protein and mRNA were detected in skeletal muscle tissues of these same animals. In null lungs, myogenin expression of RNA and protein was absent in lungs and limb muscle regions at E12, E14, E17, and E20 (data not all shown).

View larger version (40K): [in this window] [in a new window]   View larger version (125K): [in this window] [in a new window]   Figure 3.   RT-PCR of wild-type mouse (E14). (A) Myogenin mRNA appeared in leg, but not in heart, lung, or liver, of E14 wild-type (+/+) and heterozygous (+/) pups. Extracted RNA was amplified by RT-PCR (upper panel ). DNA molecular weight markers are shown in leftmost lane (100-bp ladder) and rightmost lane (D-15). Lane w is water control. Positive control of full-length cDNA for myogenin is shown in lane EMSV-cDNA myogenin. Lack of genomic DNA contamination in RNA samples is shown in the lane depicting genomic DNA. PCR product includes a 1-kb intron. Equivalent loading of RNA among samples is shown in the lower panel for the RT-PCR of S16 ribosomal protein. (B) Horse-radish brown immunoperoxidase staining of antimyogenin monoclonal F5D antibody shows presence of myogenin protein in back skeletal muscles (upper panel ), but not in lung (lower panel ) in E12 wild-type mice (same observations were found in E14 and E17 wild-type mice; data not shown). Nuclei were counterstained blue with hematoxylin. Distance of bar is 50 µm.

Lack of Skeletal Muscle Development Surrounding Pulmonary Tissues

Previous reports of the absence of skeletal muscle fibers in myogenin null mice (27) were confirmed. No diaphragmatic hernias were noted in myogenin null mice. This is important because congenital diaphragmatic hernias are associated with lung hypoplasia (45). The entire thoracic regions of six myogenin null neonates were sectioned. The diaphragm was always intact across the abdominal cavity albeit thin and fibrous (27). However, the diaphragm of the myogenin null mouse never demonstrated visceral herniation into the intrathoracic cavity, and amniotic fluid volume appeared normal, which suggests that oligohydramnios did not occur. Both of these features are essential to distinguish if fetal breathing movements alone can cause pulmonary hypoplasia. With H&E staining, laryngeal morphology of E20 mice pups revealed drastically reduced skeletal muscle fiber formation in the myogenin null versus wild-type littermates (Figure 4). Closure of the glottic rima depends on the intrinsic laryngeal muscles, which are absent in the null mouse but readily seen in the wild-type littermate.

View larger version (122K): [in this window] [in a new window]   View larger version (120K): [in this window] [in a new window]   Figure 4.   The almost complete absence of laryngeal skeletal musculature (see arrow indicating skeletal muscle region) in myogenin null mice (A) was associated with a failure of the esophagus to maintain its oval appearance as found in wild-type (B) littermates at E20. Cross-sections were taken from similar regions of the larynx and stained with H&E. We speculate that lack of skeletal muscle in the laryngeal areas impairs the normal closure of the epiglottic, which enables the back-pressure or stenting forces of lung fluid during development. Laryngeal cartilaginous structures are grossly similar in wild-type and null, but the markedly reduced skeletal musculature seen in the myogenin null larynx suggests an inability to adequately close the glottic opening. Distance of bar is 500 µm.

myogenin Null Mouse Embryos Have Lung Cells That Express SP-A Protein and SP-C mRNA

Widespread lumenal surface and cell-specific SP-A immunoreactivity were found in wild-type littermates in contrast to localized SP-A immunoreactivity overlying specific cells, presumably type II pneumocytes or possible Clara cells, in myogenin null lung cells at E17 to E20 (Figures 5A and 5B). No SP-A immunoreactivity was detected in E14 null or wild-type lungs (data not shown). The finding of SP-A immunoreactivity (a marker of type II pneumocyte and possible Clara cell differentiation [15]) in the myogenin null and wild-type littermate lungs suggests that these early lung cells are differentiating (46). The more distal epithelial cells expressed the highest immunoreactivity of SP-A in wild-type lungs than in null lungs. Whereas fetal lung growth and size appear to depend on normal skeletal muscle fiber formation, skeletal muscle development does not impair type II pneumocyte and possible Clara cell differentiation in early lung development.

View larger version (134K): [in this window] [in a new window]   View larger version (41K): [in this window] [in a new window]   View larger version (110K): [in this window] [in a new window]   Figure 5.   SP-A is present in lung sections from both heterozygous (A) littermates of myogenin null (B) pups at E19. Antihuman SP-A antibody was exposed to the secondary antibody immunoperoxidase (reddish-brown color). Sections were counterstained with hemaoxylin (bluish color; Zymed Histostain Kit). Distance of bar is 20 µm. (C) SP-C mRNA was present in RNA extracted from lungs of both heterozygous (+/) littermates and myogenin null (/) lungs of E14 mice. SP-C RNA primers were used in RT-PCR amplification. Negative controls are shown in water with RT-PCR reagents minus lung RNA (lane RT-W) and water only (lane W). S16 ribosomal primers revealed equivalent loading and integrity of RNA samples (data not shown). (D) In lumenal spaces of both wild-type (+/) and null lungs (/), the presence of lamellar bodies (LB) and tubular myelin (TM) are suggestive that the null lung cells are able to differentiate and to participate in surfactant synthesis to a certain extent, although not as mature as wild-type lungs. Null lungs have increased density of glycogen granules (GG) compared with wild-type lungs. Original magnification ×2,000. Bar = 2 µm.

To assess for onset of early lung cell differentiation, we performed RT-PCR analysis for SP-C expression with whole lungs dissected from E14 null and wild-type embryos. We found SP-C expression, an early marker for lung cell(s) differentiation, to be detectable in both null and wild-type lungs from embryos as early as E14 (Figure 5C). At this early stage, we could not detect any expression of SP-A mRNA, although S16 ribosomal mRNA could readily be detected (data not shown), consistent with previous mouse lung studies (41). This finding in myogenin null E14 embryos suggests that the activation of SP-C gene transcription, a marker of early lung cell differentiation, occurs as a result of process(es) independent of skeletal muscle development. As seen by electron microscopy (Figure 5D), the myogenin null lung had obvious lamellar bodies and tubular myelin suggestive of surfactant precursor production. Indicating less mature features, the null lungs compared with wild-type lungs at E17 and E20 had slightly increased glycogen density (29, 47).

Attenuated Lung Cell Proliferation in E14 Null Mice

To evaluate the underlying basis for the hypoplastic lungs that were grossly smaller and contained approximately 20% less total DNA at term, we investigated indices of cell proliferation through BrdU incorporation studies (Figure 6). BrdU labeling at the E14 and E17 stages suggested that nearly 50% of all lung cells in the wild-type embryos were immunoreactive, whereas BrdU incorporation in myogenin null lungs was significantly decreased with only 36% of all lung cells being immunoreactive (Table 2). PCNA immunolabeling also revealed significantly higher proliferative rates in the wild-type lungs (data not shown). Our preliminary investigation showed that wild-type lungs at E14 and E17 stages had markedly increased proliferation in mesenchymal compartments (45 to 61%) versus epithelial (bronchial) areas (24 to 38%) (data not shown). In contrast, the null lungs at E14 and E17 had similar rates of cell proliferation (30 to 36%) in mesenchymal and epithelial compartments. At E19, wild-type and null lung proliferation markers all decreased < 8% in either cell type. It is important to appreciate the architectural differences (wild-type versus null) in E19 lungs, although both exhibit decreased proliferation. At term, the wild-type lungs continue to have canalicular morphology (see Figures 5 and 7 for comparisons).

View larger version (141K): [in this window] [in a new window]   View larger version (139K): [in this window] [in a new window]   Figure 6.   Immunoperoxidase brown staining (see arrows) of anti-BrdU in lungs from E14 heterozygous (A) littermates and myogenin null (B) mice (see Table 3 for a tabulation of these results). Nuclei were counterstained with hematoxylin (blue). Distance of bar is 20 µm. Similar results were found with anti-PCNA labeling.

View larger version (145K): [in this window] [in a new window]   View larger version (137K): [in this window] [in a new window]   View larger version (143K): [in this window] [in a new window]   View larger version (146K): [in this window] [in a new window]   Figure 7.   More Bax protein immunoreactivity seen in lungs from myogenin null (A) than in lungs from heterozygous (B) or wild-type littermates in E19 mice. Rabbit antimouse Bax monoclonal antibody was stained with a secondary antibody labeled with brown immunoperoxidase (see arrows). Distance of bar is 20 µm. TUNEL staining is present in lungs from myogenin null (C) mice as compared with their wild-type (D) littermates at E19. An avidin-peroxidase coupled antibody with brown diaminobenzidine substrate (see arrow) was used to react with TUNEL. Nuclei were counterstained blue with hematoxylin. Distance of bar is 50 µm.

Apoptosis in the myogenin Null Lung Is Present at E17 and E20

As gestation continued, lung cellular proliferation decreased (Table 2). Concurrently, markers of apoptosis (TUNEL and Bax positive) were initially absent at E14, then increased at the E17 stage in lungs of both wild-type and myogenin null mice, and remained high at E20 in the myogenin null mice as compared with wild-type littermates (Table 3). The increased cell death appeared specific to the myogenin null lungs as adjacent heart and liver in the same section did not reveal these same results. Indices of apoptosis (Bax, a well-described pro-apoptotic effector [Figures 7A and 7B] and TUNEL [Figures 7C and 7D]) were more evident in E17 myogenin null mice than in age-matched, wild-type littermates, which had little positive labeling (same data was found in E17 and E20 lungs; data not shown). Based on the location of TUNEL-positive cells, we approximate that nearly 70 to 80% of the total number of apoptotic cells seen were likely bronchial or mesenchymal cells, not epithelial cells. These findings appeared specific for apoptotic cells as we could also easily visualize positive Bax or TUNEL labeling in skin (data not shown), consistent with previous studies (36, 40). Negative controls without primary antibodies showed no labeling above background. We considered the possibility that humoral cells, e.g., leukocytes, may be responsible for the apoptotic findings in the lung, but with high-power magnification, we were able to find a large proportion of TUNEL-positively or BAX-positively labeled lung cells outside of blood vessels. By using electron microscopy, we were able to see some lung cells demonstrating features of apoptosis, including nuclear condensation, chromatin margination and clumping on nuclear membranes, cell shrinkage, and membrane blebbing without evidence of inflammation (Figure 5D). Despite our best efforts, the null lungs appeared to have a greater sensitivity to necrosis and fixation artifact. At term, the cells found to be TUNEL or Bax positive were predominantly bronchial cells, with a minor proportion as mesenchymal cells.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Knockout of a specific gene for skeletal muscle differentiation, myogenin, leads to the predicted phenotypic defect in skeletal muscle, i.e., lack of normal muscle fiber development (27). However, a novel observation of the current study is that the knockout of this same gene also leads to abnormal lung morphology during development. As myogenin is only expressed in skeletal muscle, not in lung (Figure 3), we postulate that lung hypoplasia is secondary to abnormal development of skeletal muscle. Clinical evidence supports this idea. For example, severe congenital myopathies in humans are often associated with pulmonary hypoplasia (7).

The lack of myogenin has previously been shown to affect the development of tissues other than skeletal muscle in the mouse. For example, in the myogenin null mouse, hypoplastic tibial tuberosities and other hypoplastic muscular attachments in the skeletal system have been demonstrated (27), and a moderate effect of reduced skeletal muscle formation on numbers of -motor neurons in the developing spinal cord has been shown (48). These observations, as well as the current findings, emphasize the advantage of transgenic studies to provide an in vivo opportunity to observe secondary effects of tissue-specific gene expression on other cell types. Because myogenin null mouse embryos develop without surgical trauma, there is no expectation of elevated cortisol concentrations, which can influence pneumocyte differentiation. Another advantage of the myogenin knockout mouse model is that wild-type littermates (i.e., control mice) and null mice from the same mother are exposed to the same uterine and placental perfusion environment, unlike models that use pharmacologic (e.g., nitrofen, curare, bungarotoxin) or surgical perturbations that compare different litters (45). Because myogenin is expressed exclusively in skeletal muscle (25, 49), secondary effects of deficient skeletal musculature on early mouse lung development can be studied in myogenin null mice. It would be interesting to see if proliferation and apoptosis would be similarly affected when lungs from myogenin null and wild-type mice are isolated from embryos and grown in culture.

Lack of functional skeletal muscle in utero could affect lung development by at least three different possibilities. First, the myogenin null mouse has an inadequate intrathoracic space for the developing lung, similar to chondrodystrophic mice (33). The lack of skeletal muscle produces a curved kyphotic vertebral column (27, Figure 1), narrowed intercostal spaces (data not shown), and thin diaphragm (27) that cannot increase intrathoracic space.

Second, the absence of breathing movements could contribute to insufficient mechanical stretch or strain of the developing lung and affect proliferation (56). When lung cells are stretched in vitro, their proliferation increases (11, 12, 17), possibly via platelet-derived growth factor as shown with in vitro mechanical stretch (10). The current study extends these observations to an in vivo model as decreased lung cell proliferation was noted in the absence of breathing movements in the E10 to E14 myogenin null mice (Table 2). Fetal breathing is thought to play an indirect role in facilitating an appropriate mechanical environment for normal lung growth by a normal net efflux of fluids (4, 6, 7, 18, 50). Fetal breathing would mechanically stimulate the lung by facilitating fluid pressure changes (2), and we speculate that this process is absent in the myogenin null lung.

Third, the lack of functional laryngeal, lateral cricoarytenoid, and thryoarytenoid muscles in the myogenin null mouse would lead to a chronically dilated glottis during gestation. Thus, the normal stenting effect from the occlusion of lung fluid efflux to promote formation of acini, canals, and bronchioles (5) does not likely occur in the myogenin null mouse. Tracheal ligation studies describe retention of lung liquid secretions and augmented promotion of lung saccule development (18, 51). For example, human autopsy cases of infants with tracheal agenesis are associated with pulmonary hyperplasia (2, 57). Carefully performed tracheal ligation experiments in sheep and rabbits (2) that did not cause profound oligohydramnios caused pulmonary hyperplasia. A net inward flow of amniotic fluid does not occur in normal embryos since radioopaque dyes injected into amniotic sacs of sheep and rabbits were only found in the gastrointestinal tract, not the lungs (6, 50). In fact, there appears to be a net efflux as the pulmonary parenchyma secretes a measurable volume of fluid for contributing to the amniotic volume (18). We speculate that the myogenin null mouse with no skeletal muscle to close the epiglottis does not retain as much fluid as the wild-type lung and thus experiences less fluid pressure changes in utero.

There appear to be at least two independent pathologic molecular responses in the lungs of myogenin null mouse: decreased proliferation (E10-E14) and increased apoptosis (E17-E20), both which could contribute to pulmonary hypoplasia. Mechanical strain has been reported to influence lung cell proliferation rates in vitro (11, 12) and may alter the BrdU incorporation rate in developing lungs. An aberrant mechanical environment could decrease the overall BrdU incorporation in myogenin null lungs (Table 2). Mechanoceptors and signaling pathway elements that are involved with lung tensegrity remain to be elucidated (52). With TUNEL staining, a low incidence (two per 1,000 cells) of lung apoptosis was observed in wild-type lungs of mice, which are similar to the TUNEL stain findings in normal-developing rat lungs by Scavo and associates (53). Markers for apoptosis (TUNEL and Bax [Figure 6]) indicate that cells in myogenin null lungs are undergoing greater apoptotic events in late gestation, as compared with wild-type littermates. This process may represent a default pathway in cells that are not able to develop with appropriate cues, factors, or environment.

As SP-C mRNA is a specific marker for the lung epithelium (46), its appearance in the lung cells of myogenin null mice (Figure 5) indicates that some differentiation marker genes in early lung cells are expressed in the absence of functional skeletal muscle. Thus, expression of these surfactant genes appears to occur in the absence of normal skeletal muscle development. It is known that pulmonary precursor cells give rise to nearly 40 different cell types in the mature lung (47). Transcription factors and growth factors that are expressed specifically in lung primordium are now being characterized (23, 54) and the findings may lead to additional answers of stimuli for lung cell development.

Previously the death of myogenin null mice upon birth was attributed to the lack of breathing (27), but we believe their hypoplastic lungs are noteworthy. We speculate that if myogenin null mice could be mechanically ventilated at birth, their underdeveloped lungs would have decreased total lung capacities and be incompatible with extrauterine life. Our current observations raise the clinical concern that infants born with defective skeletal muscles may have comorbid features of lung hypoplasia. Gas exchange requires adequate surface area of the pulmonary parenchyma, and small lungs, even if adequately mature, are more liable for complications. Prenatal assessment of absent or hypoactive fetal movements may accurately identify infants at risk for lung hypoplasia, which is a predictor of high morbidity and mortality (7, 8, 14, 16). Thus, mechanical ventilation and exogenous surfactant may be insufficient to overcome the immature lung architecture and incipient apoptosis in such infants with lung hypoplasia. We speculate that future management of such cases might include factors to accelerate terminal sac formation and anti-apoptotic therapies, e.g., caffeine, calpain, caspase inhibitors, insulin-like growth factor-1, verapamil, etc., to counter programmed cell death.

Our findings indicate that the phenotype of lungs and skeletal muscles in mice without myogenin is very similar to mice lacking both p21^CIP and p57^KIP2, as reported by Zhang and coworkers (55) in 1999. p21^CIP and p57^KIP2 are both expressed in many tissues, including lungs and skeletal muscle, but their dual omission also led to a perinatal lethal phenotype with similar aberrant pulmonary development in the face of a very similar skeletal muscle defect (55) as seen in the myogenin null mouse. p21^CIP and p57^KIP2 are not skeletal muscle-specific or lung tissue-specific cell cycle mediators, so it is unclear why their combined genetic ablation would lead to predominant lung or skeletal muscle defects. Nevertheless, our observations disagree with a statement given by Zhang and coworkers (55) who wrote that myogenin did not have a role in lung development. Our findings suggest that myogenin and subsequent muscle fibers have an indirect, but critical, role in lung development. Because myogenin is expressed in skeletal muscle, but not in lungs, the lung defects seen in the myogenin null mice could not have been due to the absence of myogenin protein in the lung, but more likely is due to a secondary effect from the absence of skeletal muscle fibers. Further studies are warranted, but we postulate that both mouse lines demonstrate the importance of functional skeletal muscle fiber in utero for normal lung development. Further, Zhang and colleagues (55) reported no evidence of altered apoptosis in the hypoplastic lungs of the mice lacking both p21^CIP and p57^KIP2. In contrast, we observed an increased apoptosis in late embryonic stages (E17 and E20) in the myogenin null mice. Since Zhang and colleagues (55) did not show data for the age of mice, we speculate that they may have made measurements at younger embryonic ages than ours. We did not find increased apoptosis in E14 mice.

In summary, a mouse model with arrested skeletal muscle development, the myogenin null mouse has permitted further insight into the complex processes of lung development. Lack of functional skeletal muscle in utero was associated with lung hypoplasia (55). In our model, the early lung cells revealed attenuated lung cell proliferation, and later fetal stages incurred augmented apoptosis. These observations emphasize that the complexity of lung development in utero depends on multiple factors, including appropriate mechanical influences from skeletal muscle. Thus, normal skeletal muscle has secondary effects on lung maturation that may have clinical implications for the management of neonates with muscle-nerve disorders and other embryologic defects producing lung hypoplasia.