Apparent normal lung architecture in protein tyrosine phosphatase-ς-deficient mice

Jane Batt, Ernest Cutz, Chris Fladd, Daniela Rotin


Protein tyrosine phosphatase-ς (PTP-ς) is a member of the mammalian LAR family of phosphatases, which is characterized by a cell adhesion-like ectodomain, a single transmembrane segment, and two tandemly repeated intracellular catalytic domains. The expression of PTP-ς is developmentally regulated in epithelial, neuronal, and neuroendocrine tissues. We previously showed that PTP-ς is strongly expressed within the fetal, but not adult, rat lung and is localized to the Clara cells and type II pneumocytes. In view of the developmentally regulated pulmonary expression of PTP-ς, we performed a detailed histological and ultrastructural study of the lungs of PTP-ς knockout mice we have generated. Our findings indicate no apparent structural abnormalities in the lungs of PTP-ς−/− mice, including airway and alveolar epithelium. In addition, pulmonary neuroendocrine cells also appear normal, in contrast to pituitary, pancreatic, and gastrointestinal endocrine cells, in the knockout mice, suggesting different developmental regulation of these neuroendocrine cells. These observations suggest compensation for the absence of PTP-ς during development by related family member phosphatases, such as LAR.

  • pulmonary
  • knockout mice
  • neuroepithelial body

protein tyrosine phosphatase (PTP)-ς is a transmembrane receptor PTP and a member of the mammalian LAR family, which consists of the three closely related enzymes: PTP-ς, PTP-δ, and LAR. PTP-ς contains a cell adhesion molecule (CAM)-like ectodomain, a single transmembrane domain, and two tandemly repeated intracellular phosphatase domains (13,19, 22, 25, 27). The proximal phosphatase domain is catalytically active, whereas the second domain is inactive and serves a regulatory role (21). The CAM-like ectodomain of PTP-ς bears a strong resemblance to the ectodomain of neural CAMs of the Ig superfamily, such as L1 and N-CAM (3, 6). These molecules regulate the development of specific axonal projections in the nervous system. The substrate and ligand for PTP-ς and the signaling networks in which it functions remain unknown.

PTP-ς expression is tightly controlled and developmentally regulated within epithelial, neuronal, and neuroendocrine tissues and organs (13, 15, 23, 25). It plays a critical role in the development of the pituitary, pancreas, enteroendocrine gut, and peripheral nerve. PTP-ς knockout mice demonstrate pituitary and pancreatic islet hypoplasia associated with deficiencies of growth hormone (GH), prolactin, and insulin (1, 4, 20). Severe GH deficiency contributes to neonatal hypoglycemia in the PTP-ς−/− mice and a high neonatal mortality rate (60%) (1). The peripheral nerve in the PTP-ς-deficient mice shows developmental delay (20) and significant abnormalities of axon guidance and regenerative capacity after injury (12).

Within the rat lung, PTP-ς is predominantly and strongly expressed in proliferating and differentiating epithelial cells (fetal Clara cells and type II pneumocytes) of the embryonic, fetal, and neonatal airway and alveolar sacs (9). PTP-ς is not found in quiescent adolescent and adult rat pulmonary epithelium. On the basis of this cell-specific and developmentally regulated pulmonary expression and the absolute requirement for PTP-ς in the developing nervous and neuroendocrine systems, we hypothesized that PTP-ς would also play a role in the development of the mammalian respiratory system. We thus performed detailed histological and ultrastructural assessment of the lungs in PTP-ς-deficient mice.


The full-length PTP-ς gene was inactivated, and mice were characterized as previously described (20). The PTP-ς+/− mice are phenotypically indistinguishable from PTP-ς+/+ mice. The PTP-ς−/− animals consisted of three cohorts: most (60%) died as neonates within hours of birth, 37.5% demonstrated growth retardation and succumbed to a wasting syndrome by 2–3 wk of age, and 2.5% survived to adulthood, appeared healthy, and were fertile.

All animal experimentation was conducted in accordance with accepted standards of humane animal care.

Lac Z staining.

The knockout cassette used to generate the PTP-ς knockout mice contains the β-galactosidase gene, allowing Lac Z staining to determine the expression pattern of PTP-ς. Animals were killed, and the lungs were harvested and fixed in 0.1 M sodium phosphate buffer (pH 7.9) containing 1% formaldehyde, 0.1% glutaraldehyde, 2 mM MgCl2, and 5 mM EGTA for 6 h. Organs were washed for 2 h with four exchanges of a wash buffer (2 mM MgCl2, 0.01% deoxycholate, and 0.02% NP-40 in 0.1 M sodium phosphate buffer, pH 7.9) at room temperature. The tissue was then incubated in PBS containing 5 mM ferricyanide, 5 mM ferrocyanide, 2 mM MgCl2, and X-gal (Roche; 0.1 mg/ml) overnight at 37°C. Tissues were subsequently rinsed in 70% ethanol, embedded in paraffin, and sectioned.

Lung wet-to-dry weight ratios.

Newborn litters were killed via decapitation, and the right lung was immediately excised, gently blotted on a cotton towel, and weighed to obtain the lung wet weight. The lungs were then placed in a 50°C oven and weighed daily until a stable dry weight was obtained (≥72 h). Wet-to-dry lung weight ratios were determined for PTP-ς−/−, PTP-ς+/−, and PTP-ς+/+ newborns. An increased ratio is indicative of an increase in intravascular and/or interstitial lung water content.


Representative mice from each of the three PTP-ς−/−cohorts and sibling controls were euthanized. Newborn mice were decapitated, and the lungs were dissected and fixed overnight in 4% paraformaldehyde at 4°C. Adult and 2- to 3-wk-old mice were sedated with an intraperitoneal injection of chloral hydrate and subsequently perfusion fixed with 4% paraformaldehyde. Alternatively, these two cohorts of mice were euthanized by lethal intraperitoneal injection of pentobarbital sodium. The lungs were then dissected and fixed overnight in 10% buffered formalin at 4°C. All lung tissue was subsequently embedded in paraffin, sectioned (5 μm), and stained with hematoxylin and eosin (H & E) or periodic acid-Schiff (PAS).

Electron microscopy.

Representative newborn, 2- to 3-wk-old, and adult PTP-ς−/− mice and appropriate controls were euthanized; the lungs were dissected and fixed overnight in 1% glutaraldehyde-4% formaldehyde in 0.1 M phosphate buffer. Lungs were washed once in 0.1 M phosphate buffer for 5 min and then postfixed in 2% OsO4. Reduced osmium fixation (2% OsO4 and ferrocyanide) was used to demonstrate cytoplasmic glycogen. Sections (1 μm) were dehydrated in graded acetone, embedded in Epon, and stained with toluidine blue; ultrathin sections were stained with uranyl acetate and lead citrate. Electron-microscopic examination was performed with a transmission electron microscope (model 201, Phillips).


Immunohistochemical localization of calcitonin gene-related peptide (CGRP), a marker of neuroepithelial bodies (NEB) in rodent lungs, was carried out as previously reported (5). Briefly, representative mice from each of the three PTP-ς−/−cohorts were euthanized, and the lungs were dissected and fixed for 30 min (newborn) or 2 h (2–3 wk old and adult) at room temperature in Bouin's fixative. Lungs were subsequently washed four times with PBS overnight at 4°C, embedded in paraffin, and sectioned (5 μm). Sections were blocked and incubated with a rabbit polyclonal CGRP antiserum (Chemicon) at a dilution of 1:400 overnight at 4°C, followed by biotinylated secondary antibody, streptavidin-HRP (Vectastain Elite ABC kit; Vector Labs) and developed in diaminobenzidine (0.5 mg/ml) and H2O2 (0.03%). Appropriate positive and negative (omission of primary antibody) controls were performed.


Developmental expression of PTP-ς in mouse lungs.

The PTP-ς knockout cassette contains the β-galatosidase gene (20). Lac Z staining was therefore undertaken to localize the expression of PTP-ς within the lungs of the three cohorts of PTP-ς−/−, PTP-ς+/−, and PTP-ς+/+ mice. PTP-ς is expressed within the walls of the alveolar sacs, airway epithelium, and pulmonary vessels of newborn mice (Fig. 1, A andB). At 2–3 wk of age, LacZ staining is evident in the alveolar walls and pulmonary vessels but is no longer expressed in the airway epithelium (Fig. 1, D and E). As in other tissues, PTP-ς expression decreases dramatically as the animal matures and is absent from the lungs of the adult mice (Fig. 1,G and H). These results are in agreement with our previously observed decline of PTP-ς expression with maturation of rat lung epithelia (9).

Fig. 1.

Lac Z staining of protein tyrosine phosphatase (PTP)-ς−/−, PTP-ς+/−, and PTP-ς+/+ newborn (NB), 2- to 3-wk-old, and adult lungs. As the β-galactosidase gene replaces PTP-ς in the knockout cassette, blue Lac Z staining represents the PTP-ς expression pattern. The pattern of Lac Z-positive staining in PTP-ς−/− mice (A, D, and G) parallels that in PTP-ς+/− mice (B, E, andH). Lac Z staining is evident in alveolar sacs, airway epithelium, and pulmonary vasculature of newborn mice (A andB). Lac Z staining is still evident in alveoli but absent from airway epithelium by 2–3 wk of age (D andE). Lac Z is no longer expressed in adult lung (Gand H). PTP-ς+/+ mice act as negative controls for Lac Z staining (C, F, and I). Scale bar, 40 μm. Insets depict higher magnification of segments of the tissue.

Analysis of pulmonary edema.

During the original phenotypic characterization of the PTP-ς−/− mice, pulmonary edema was noted in some of the newborn animals (20). To further investigate this observation and analyze its prevalence, the lung wet-to-dry weight ratio was determined for a large series of newborn mice. The presence of fluid within the lung manifests as an increase in the lung wet-to-dry weight ratio. Figure 2 shows no significant difference in the lung wet-to-dry weight ratio between PTP-ς−/−, PTP-ς+/−, and PTP-ς+/+ newborn mice. This suggests that, overall, there is no significant pulmonary edema in the PTP-ς−/−newborn.

Fig. 2.

Lung wet-to-dry weight ratios of newborn mice. An increase in lung wet-to-dry weight ratio indicates alveolar fluid or an increase in intravascular volume. There is no significant difference between lung wet-to-dry weight ratios of PTP-ς−/−(n = 20) and sibling PTP-ς+/−(n = 43) and PTP-ς+/+ (n= 21) newborn mice.

Analysis of lung architecture.

Analysis of H & E-stained sections of lungs from newborn and adult PTP-ς−/− mice revealed no apparent abnormalities (Fig.3, A and E). Specifically, the alveoli and epithelial lining of the bronchioles appeared normal, and alveolar pulmonary edema was not evident in the newborn PTP-ς−/− mice (Fig. 3 A). One 18-day-old PTP-ς−/− mouse (2- to 3-wk-old cohort) demonstrated significant bronchial epithelial hyperplasia and possibly metaplasia (data not shown), but similar findings were not evident, and the lungs appeared normal in the large majority of this cohort (Fig.3 C).

Fig. 3.

Hematoxylin-and-eosin-stained sections of PTP-ς−/− newborn (A), 2- to 3-wk-old (C), and adult (E) mice and PTP-ς+/+ newborn (B), 2- to 3-wk-old (D), and adult control mice (F). Scale bar, 100 μm. All 3 cohorts of PTP-ς−/− mice demonstrate normal histology; i.e., there are no differences in appearance of alveolar sac (AS), alveoli (AL), or bronchioles (BR) between PTP-ς−/− mice and age-matched PTP-ς+/+controls.

The Clara cells of the airways of newborn mice contain a significant amount of cytoplasmic glycogen that is rapidly metabolized in the first 24–48 h after birth (11). Because PTP-ς is expressed in the Clara cell, we used PAS staining with diastase digestion to visualize the glycogen stores of newborn PTP-ς−/− animals. Our results show no difference in the distribution of glycogen or the duration of positive glycogen staining in the PTP-ς−/− newborn mice (Fig.4, A and B) compared with PTP-ς+/− and PTP-ς+/+sibling controls (Fig. 4, C and D). Thus, although the PTP-ς knockout mice exhibit a delay in several aspects of development, glycogen metabolism in Clara cells at birth is not altered.

Fig. 4.

Periodic acid-Schiff staining of newborn PTP-ς−/− mice (A) and sibling PTP-ς+/+ mice (C) reveals magenta-colored neutral mucopolysaccharides in epithelial cells of the airway representing Clara cells (arrows). Pretreatment with diastase (B and D) reveals loss of magenta staining (arrowheads), confirming presence of glycogen. No differences are noted between PTP-ς−/− and PTP-ς+/+ newborns. Scale bar, 40 μm.

Electron microscopy of lungs from all three cohorts of the PTP-ς−/− mice was undertaken to assess subcellular components of the cell types expressing PTP-ς, such as the lamellar bodies of the type II pneumocytes and the secretory granules of the Clara cells. The Clara cells of newborn, 2- to 3-wk-old, and adult PTP-ς−/− mice (Fig.5 A, C, and E) appear normal and indistinguishable from Clara cells of the PTP-ς+/− or PTP-ς+/+ control mice (Fig. 5, B, D, and F). The Clara cells of the newborn mouse contain significant cytoplasmic glycogen with few or no secretory granules. Postnatally there is an expected significant decrease in cytoplasmic glycogen content, with a corresponding increase in the cell organelles, including secretory granules, and a modest increase in the mitochondrial content (11). The cell organelles of type II pneumocytes, including their lamellar bodies, also appeared normal in all three cohorts of PTP-ς−/− mice and indistinguishable from the type II cells of the PTP-ς+/− and PTP-ς+/+ sibling control mice (Fig. 6).

Fig. 5.

Electron micrographs of airway epithelium in PTP-ς−/− newborn (A), 2- to 3-wk-old (C), and adult (E) mice and PTP-ς+/+ newborn (B), 2- to 3-wk-old (D), and adult (F) mice. Magnification ×3,572. Cytoplasm of Clara cells in lungs of newborn mice contain large amounts of cytoplasmic glycogen (A and B) appearing as electron-dense material (GLY) with reduced osmium fixation (inset). Clara cell glycogen content decreases and secretory granules (arrowhead) and mitochondria (arrow) increase as the cell matures (C–F). Clara cells and ciliated cells of PTP-ς−/− mice are indistinguishable from PTP-ς+/+ mice.

Fig. 6.

Electron micrograph of a type II pneumocyte from a 2- to 3-wk-old PTP-ς−/− mouse (A) and a sibling PTP-ς+/+ mouse (B). Magnification ×11,324. Well-formed lamellar bodies (L) are evident in both micrographs. Type II cells from all 3 cohorts of PTP- ς−/− mice appeared normal.

Pulmonary neuroendocrine cells.

We previously described abnormalities of the neuroendocrine system (pituitary, pancreas, and enteroendocrine cells of the gut) of the PTP-ς−/− mice (1, 20). To determine whether the pulmonary neuroendocrine cell (PNEC) system may be similarly affected by the loss of PTP-ς, we immunostained the lungs of the PTP-ς−/− mice for CGRP, a marker of PNEC and NEB (5). Our findings show that PNEC and NEB in the lungs of PTP-ς−/− mice (Fig. 7,A and B) were comparable to those in the lungs of PTP-ς+/− and PTP-ς+/+ sibling control mice (Fig. 7, C and D).

Fig. 7.

Positive immunostaining of neuroepithelial bodies (NEB) for calcitonin gene-related peptide (arrows) in a 2- to 3-wk-old PTP-ς−/− mouse (A and B) and a PTP-ς+/+ sibling mouse (C and D). Calcitonin gene-related peptide-positive staining appears brown. NEB are normally sparse and appear as small clusters of intraepithelial cells preferentially located at the airway bifurcation. NEB in PTP-ς−/− mouse are similar to those in control PTP-ς+/+ mouse. Scale bars, 100 μm (A andC) and 20 μm (B andD).


We previously demonstrated by in situ hybridization and Northern blot analysis that PTP-ς expression is developmentally regulated in the rat lung (9, 14). The highest levels of expression of PTP-ς in lungs were evident in fetal and newborn epithelium of the bronchi, terminal bronchioles, budding air sacs, and alveoli, with significant downregulation of expression as the rat matures. Parallel immunostaining for surfactant proteins A and B, proliferating cell nuclear antigen, and [3H]thymidine uptake studies revealed that PTP-ς was predominantly expressed in the undifferentiated and proliferating fetal alveolar type II cells and fetal Clara cells of the airways (9). Fetal type II cells are the precursors for adult type II cells (10). Fetal Clara cells are the stem cells of the small conducting airway epithelium and give rise to adult Clara cells and ciliated cells (10). Once differentiation was complete and proliferation was reduced to the low level of epithelial cell turnover normally seen in the quiescent healthy postnatal lung, PTP-ς expression became undetectable (9).

In the present study, the pulmonary PTP-ς expression pattern was determined by Lac Z staining of the PTP-ς−/− and PTP-ς+/− mice. In agreement with our previous in situ hybridization results in the rat lung (9), Lac Z staining demonstrates expression of PTP-ς in the airways and respiratory epithelia of newborn mice; moreover, this expression is developmentally downregulated. Although we previously failed to detect PTP-ς expression in the rat pulmonary vasculature using in situ hybridization (9), Lac Z staining (Fig. 1) clearly demonstrates significant expression of PTP-ς in the blood vessels (smooth muscle and, possibly, endothelium). The reason for the lack of previous detection of PTP-ς in the vasculature is not known but may stem from the different methodology employed, which may be less sensitive.

During the initial phenotypic analytic screen of the PTP-ς−/− mice, the occasional neonate was found to have pulmonary edema (20). Subsequent systematic collection and assessment of a large number of newborn PTP-ς−/− mice, however, demonstrated that this was not a general phenomenon. H & E analysis of PTP-ς−/− newborn lungs did not reveal evidence of alveolar fluid (Fig. 3 A). Wet-to-dry lung weight ratios were similar in PTP-ς−/−, PTP-ς+/−, and PTP-ς+/+ newborns, arguing against the presence of alveolar fluid/pulmonary edema in the knockout neonate (Fig. 2). On the basis of our recent and present work, we conclude that the high neonatal mortality rate of the PTP-ς−/− mice is the result of hypoglycemia, which at least in part results from GH deficiency, and is not due to pulmonary edema (1).

Airways of the PTP-ς−/− mice appeared normal on H & E-stained sections and indistinguishable from the airways of PTP-ς+/− and PTP-ς+/+ mice (Fig. 3). As expected, PAS staining and diastase digestion revealed glycogen in cells of the airways of PTP-ς−/−, PTP-ς+/−, and PTP-ς+/+ newborn mice (Fig.4). Many cell types of the neonatal lung, including Clara cells of the airway and alveolar type II cells, have significant glycogen stores (11). In the rat Clara cell, glycogen is immediately and rapidly metabolized postpartum, so that within 48 h of birth there is a massive decrease in the volume density (fraction of cell volume) of glycogen. This is followed by a continual, but very gradual, decrease of glycogen into adulthood. Glycogen mobilization at birth is presumably important for the maturation of the secretory apparatus of the Clara cell. In addition, because Clara cells are the progenitors for the cells of the terminal bronchioles, glycogenolysis in Clara cells may be critical for the proliferative burst they undergo during the first 48 h after birth. Glycogen deposition and mobilization, as assessed by PAS staining and diastase digestion, were similar between PTP-ς−/− and PTP-ς+/+ mice.

Ultrastructural analysis of different pulmonary epithelial cells, including alveolar type II cells, Clara cells, and ciliated cells, did not reveal any significant differences between the PTP-ς−/− and PTP-ς+/− or the PTP-ς+/+ mice (Figs. 5 and 6). The Clara cell population demonstrated typical postnatal maturation changes (11) when assessed in the three cohorts of animals. The type II cells and ciliated cells of the airway in all three cohorts of the PTP-ς−/− mice also appeared ultrastructurally normal.

We recently showed that the PTP-ς−/− mouse demonstrates significant abnormalities of the endocrine pancreas, pituitary, and enteroendocrine cells of the gut (1). Here we assessed the pulmonary neuroendocrine system by immunostaining for CGRP, a predominant neuropeptide of mouse NEB. The NEB appeared normal in the PTP-ς−/− mice and indistinguishable from the immunostaining pattern in the PTP-ς+/− and PTP-ς+/+ sibling control mice. In knockout mice lacking mammalian achaete-scute homolog-1, a mammalian homolog of theDrosophila achaete-scute genes, pulmonary NEB are completely absent, but the pancreatic islets and enteroendocrine cells in the gut develop and appear normal (2). These contrasting phenotypes of the mammalian achaete-scute homolog-1 and PTP-ς knockout mice suggest that the development and differentiation of the neuroendocrine cell component of different endodermal derivatives (i.e., gut, pancreas, and lung) are under tissue-specific regulatory control.

Although pulmonary mechanics were not formally assessed in this study, no obvious abnormalities of ventilation were evident during routine activity in the PTP-ς−/− animals. Specifically, they were not found to be tachypneic, nor was there any evidence of increased respiratory effort.

In view of the tightly controlled and developmentally regulated expression of PTP-ς within the lung, it was a surprise to see normal lung architectural development in the PTP-ς-deficient mouse. This is particularly so when one considers that PTP-ς is essential to the development of other organs (e.g., central nervous system, peripheral nervous system, pituitary), where a similar temporal expression pattern is seen, and that mice deficient in receptors bearing a marked resemblance to PTP-ς demonstrate a dramatic lung phenotype. For example, Dutt-1/Robo-1 is a mammalian transmembrane receptor possessing an ectodomain very similar to PTP-ς and the Ig family of neural CAMs (24). The ectodomains of these proteins consist of repeats of Ig-like domains and fibronectin type III repeats. Dutt-1 is probably a tumor suppressor gene in mammals (24). It has also been identified as the human homolog of the Drosophila roundabout (Robo) (8) and has been proposed to play an essential role in the guidance and migration of axons, myoblasts, and leukocytes in vertebrates. In mice homozygous for a targeted mutation in the Dutt-1/Robo-1 gene (which eliminates the first Ig domain of the ectodomain), the lungs demonstrate mesenchyme expansion and extensive bronchial epithelial hyperplasia (24). A 60% homozygote neonatal mortality rate results from respiratory failure. We also noted significant hyperplasia in the bronchial epithelium of one 2- to 3-wk-old PTP-ς−/− mouse. However, this finding was not evident in the large majority of animals assessed.

In view of the normal lung architecture in the PTP-ς-deficient mice, one may speculate that, despite strong developmental expression of PTP-ς in the lung, either this protein does not play a significant role in pulmonary development or, more likely, lung development in the PTP-ς knockout animals is “rescued” by the compensatory actions of a closely related family member, such as LAR. Katsura and colleagues (7) assessed the developmental expression profile of LAR in the rat lung via immunohistochemistry and in situ hybridization. They demonstrated that, in the lung, LAR is exclusively expressed in the epithelium of the fetal airways and alveoli with persistent Clara cell and alveolar type II cell expression throughout life. Thus LAR and PTP-ς are expressed in the same cell types within the rat lung. In view of the similar expression patterns and extensive sequence homology between the two proteins, it is not unreasonable to speculate that LAR might rescue lung development in the PTP-ς-deficient mouse. The LAR knockout mouse also lacks a lung phenotype (16-18,26), and significant upregulation of PTP-ς mRNA was observed in the lungs of these knockout mice (W. Skarnes, personal communication). Again, in this scenario, one might speculate that rescue of the LAR-deficient mice by PTP-ς would permit normal pulmonary development.

PTP-δ, the third member of the mammalian LAR family, is also expressed in the lung. Katsura et al. (7) demonstrated by in situ hybridization that PTP-δ is expressed exclusively in the mesenchyme of the lung and that expression falls off significantly after birth. It is possible that PTP-δ also plays a role in rescue of the PTP-ς knockout mouse, because mesenchymal-epithelial interactions are critical during lung morphogenesis.

In conclusion, we have generated a PTP-ς-deficient mouse and found the lung architecture to be normal by detailed histological assessment. We speculate that the presence and activity of the closely related family members LAR and, possibly, PTP-δ rescue and normalize lung development in the absence of PTP-ς in these mice. Although lung architecture is not affected by the loss of PTP-ς, pulmonary mechanics, epithelial cell biology, and other lung functions of the PTP-ς−/− mice have not been investigated and may be affected by the loss of this phosphatase.


We thank A. Giffin for assistance with mouse colony maintenance, J. Hwong for assistance with electron microscopy, and V. Wong for performing the CGRP immunostaining.


  • This work was supported by grants from the Canadian Institutes of Health Research (CIHR) to D. Rotin. J. Batt is supported by a CIHR fellowship. D. Rotin is the recipient of a CIHR Investigator Award.

  • Address for reprint requests and other correspondence: D. Rotin, Programs in Cell Biology, The Hospital for Sick Children, 555 University Ave., Toronto, ON, Canada M5G 1X8 (E-mail:drotin{at}

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • 10.1152/ajplung.00069.2002


View Abstract