Targeted expression of the receptor tyrosine kinase RON in ... - Nature

228 downloads 0 Views 197KB Size Report
1Division of Pulmonary Sciences and Critical Care, Department of Medicine, ... Cancer Center, and Denver Health Medical Center, Denver, Colorado, CO 80204 ...
ª

Oncogene (2002) 21, 6382 – 6386 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc

Targeted expression of the receptor tyrosine kinase RON in distal lung epithelial cells results in multiple tumor formation: oncogenic potential of RON in vivo Yi-Qing Chen1, Yong-Qing Zhou2, James H Fisher1 and Ming-Hai Wang*,1 1

Division of Pulmonary Sciences and Critical Care, Department of Medicine, University of Colorado School of Medicine, CU Cancer Center, and Denver Health Medical Center, Denver, Colorado, CO 80204, USA; 2Division of Neurosurgery, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, People’s Republic of China 310003

RON, a member of the MET proto-oncogene family, has been implicated in the progression of certain epithelial cancers. The purpose of this study was to determine the oncogenic potential of RON in vivo in lung epithelial cells. Transgenic mice were established using surfactant protein C promoter to express human RON in the distal lung epithelial cells. These mice were born normal but developed multiple lung tumors with distinct morphology and growth patterns. Tumors appeared as a single mass in the lung around 2 months of age and gradually developed into multiple nodules located mostly in the peripheral portions of the lung. A transition from early adenomas to later adenocarcinomas was observed. Morphologically, tumors were characterized as cuboidal epithelial cells with a type II cell phenotype, grew along the alveolar walls, and projected into the alveolar septa. RON was highly expressed and constitutively activated in tumors. These results indicate that overexpression of human wild-type RON causes the formation of lung tumors with unique biological characteristics in vivo. This model provides opportunities to study the role of RON in the pathogenesis of lung tumors and to elucidate the mechanisms underlying this distinct lung tumor. Oncogene (2002) 21, 6382 – 6386. doi:10.1038/sj.onc. 1205783 Keywords: receptor tyrosine kinase; transgenic mice; lung tumors; type II cells

RON (Recepteur d’Origine Nantais) (Ronsin et al., 1993), also known as STK (stem cell-derived tyrosine kinase) in mice (Iwama et al., 1994), belongs to the MET proto-oncogene family, a distinct subfamily of receptor tyrosine kinase (Ronsin et al., 1993). The ligand for RON was identified as macrophage stimulating protein (MSP) (Gaudino et al., 1994; Wang et al., 1994), also

*Correspondence: M-H Wang, Department of Medicine, UCHSC/ Denver Health Medical Center, 777 Bannock Street, Mail Box #4000, Denver, Colorado, CO 80204, USA. E-mail: [email protected] Received 29 June 2001; revised 6 June 2002; accepted 18 June 2002

known as hepatocyte growth factor-like protein (Han et al., 1991). MSP is a serum protein originally discovered through its effects on mouse macrophages (Skeel et al., 1991). Expression of RON is observed primarily in cells of epithelial origin including lung and colon (Gaudino et al., 1995; Ronsin et al., 1993). Disruption of the RON gene (knockout) leads to the death of mouse embryos in the early stages (Muraoka et al., 1999), suggesting that RON is developmentally required. The RON gene is normally transcribed in the lungs as evident in Northern blot analysis (Ronsin et al., 1993). Immunohistochemical staining (IHCS) confirmed that RON expression is restricted to bronchial epithelial cells, but not in alveolar cells including type II cells (Sakamoto et al., 1997). Increased RON expression in the lungs was observed after administration of chemical carcinogens (Willett et al., 1997). Altered RON expression has also been found in lung cancer cell lines (Willett et al., 1998). Recently, a mutation in the RON gene has been identified in a primary human lung cancer (Angeloni et al., 2000). These results suggest that RON might be involved in the progression of certain lung tumors in vivo. To study oncogenic potentials of RON in vivo, particularly in the lungs, we decided to express human wild-type (wt) RON in the mouse lungs using transgenic techniques. The RON cDNA (Ronsin et al., 1993) was inserted into the vector 3.7SPC/SV40 (Glasser et al., 1990) between the 3.7 kb human surfactant protein C promoter (SPCp) and the 0.4 kb SV40 sT intro-poly A sequence. The transgene fragment was injected into the pronuclei of fertilized eggs isolated from B6C3/F1 hybrid mice (C57BL/66C3H, Taconic). Eggs surviving microinjection were implanted into the oviducts of pseudopregnant foster mothers (CD1, Charles River Laboratories) following previously described procedures with modifications (Hogan et al., 1996; Ho et al., 1998). The integration of the SPCp-RON transgene into the genome was determined by Southern blot analysis using mouse-tail DNA. Five founder mice, #35, 36, 39, 40, and 71, were found to harbor the transgene with varying copy number (Table 1). Founder mouse #39 died accidentally at the age of 4 months. The remaining four founder mice (lines #35, 36, 40, and 71) were bred with B6C3/F1 hybrid mice to produce offspring (F1).

Lung tumors in RON transgenic mice Y-Q Chen et al

6383

The F1 mice with the integrated transgene were determined by PCR analysis using tail DNA as templates. Germline transmission was identified in lines #35, 36 and 71. Founder mouse #40 failed to pass the transgene to offspring, probably due to the mosaic effect on the founder mouse, with the germline cells being free of the transgene (Wikenheiser et al., 1992; Lavigueur et al., 1989). Founder mice #35, #36, and #71 and their offspring were used for further studies. Unlike RON gene knockout mice, SPCp-RON transgenic mice were born normal without visible abnormalities. Their growth and reproduction was also normal in comparison with control mice, indicating that the targeted expression of RON in mouse lung has no effect on the overall development of mice. As expected, RON is highly expressed in the lungs of all three lines of transgenic animals as determined by Western blot analysis using antibodies specific to human RON (Table 1). By comparing the size of the RON protein from transgenic lungs with control cells that express mature RON with correct proteolytic conversion, it was determined that the transgene expresses the correctly processed wtRON protein (data not shown). The transgene was not detected in other organs such as liver, stomach, intestines, spleen, heart, brain, and kidney, indicating that the transgene is specifically expressed in the lungs. Between 6 and 14 months of age, depending on individual lines of transgenic mice, all founder mice and F1 mice (104 from 127 in line #71, 63 from 69 in line #35, and 46 from 54 in line #36) showed signs of respiratory distress. To determine the cause of the illness, lungs were removed from sacrificed mice. Histological examination was performed with hematoxylin-eosin staining of formalin-fixed, paraffinembedded sections. Lungs from control mice did not show any abnormalities (Figure 1a). However, multiple tumors with uniform morphology were observed in the lungs, although the overall structures of the lung were still intact (Figure 1b). Most of the tumors were located in peripheral portions of the lung. The majority of tumors appeared as solid-alveolar adenomas with some atypical cells showing early progression towards

Table 1 Founder mice #35 #36 #39 #40 #71

adenocarcinoma. The tumors displayed a distinctive colummar-to-cuboidal morphology (Figure 1c). They grew along the alveolar walls and projected into the alveolar septa. Other organs or tissues were also examined and no tumor formation was observed. Also, no tumors were found in littermate controls (more than 160 mice examined within 8 – 18 months. These results suggest that tumor formation is not a randomly occurring event in these transgenic mice. A summary of tumors found in all three lines of transgenic mice is shown in Table 1. We next determined the growth patterns of lung tumors in SPCp-RON mice. Experiments were performed to see when tumors appeared in the lung and how long it takes for tumors to become multiples in the lung. To this end, the lungs from F1 mice (line #71) were collected at different months after the mice were born. A single tumor mass was found in the lung in three individual mice as early as two months of age. No tumors were found in newborn or one month old mice (three mice were examined in each age group). Around 4 – 6 months of age, several tumor nodules were observed. Up to 8 – 14 months, numerous tumor masses were seen. At these stages, some mice showed signs of respiratory distress. These results suggest that the tumors are most likely to be initiated two months after birth. The progression of tumors from a single mass to multiple nodules appears to be a slow process taking about 6 – 10 months. Because many alveolar cells express the RON transgene, the multiple tumors are likely to originate from different individual cells within this several month period. However, the erogenous spreading of tumors loosened from alveolar walls may also play a role. To determine the relationships between transgene expression and lung tumor formation, it is important to find if RON is expressed and activated in lung tumors. To this end, immunohistochemical staining (IHCS) was performed on lung sections using rabbit IgG antibodies specific to human RON (Chen et al., 2000). The results are shown in Figure 2. Immunoreactive RON was strongly detected in tumors (Figure 2a). The RON staining is mostly displayed as surface

Biological properties of SPCp-RON transgenic mice and their lung tumors

Transgene copy number1

Transgene transmitted to F1 mice

Expression/activation of RON in tumors2

8 3 34 13 12

Yes Yes ND4 No Yes

++/++ +/+ +++/++ 7 ++/++

Type of lung adenomas/ adenocarcinomas Alveolar, Alveolar, Alveolar, ND Alveolar,

solid solid solid solid

Incidence of lung tumors in F1 mice3 91.3% (63/69) 90.1% (81/88) ND ND 91.3% (116/127)

1

The copy numbers of the transgene was determined by Southern blot analysis using the injected DNA fragment as the positive control. 2The levels of RON expression and activation in the lungs were determined by Western blot analysis after immunoprecipitation of RON with the monoclonal antibody specific to human RON (Clone ID2) (Chen et al., 2000) using 3 mg of lung homogenate per sample. MDCK cells expressing RON were used as a positive control (++++) (Wang et al., 1994). RON phosphorylation was determined in the phospho-tyrosine assay using mAb 4G10 (Upstate Biotechnology Inc., Lake Placid, NY, USA) as the detecting antibody. The intensity of the phosphorylated RON protein was determined by comparison with those in MDCK-RON cells stimulated with 5 nM MSP (++++). 3Tumor incidence in the lungs of F1 mice was calculated by dividing the number of mice with lung tumors by total numbers of mice with the transmitted transgene. Mice were sacrificed at 6 – 20 months after their birth. 4ND: not done Oncogene

Lung tumors in RON transgenic mice Y-Q Chen et al

6384

Figure 1 Histological features of lung tumors in SPCp-RON transgenic mice. (a) Normal mouse lung. A lung section from a matched littermate control mouse (406 magnification) was used as the control. (b) Multiple lung adenomas in a lung section of a transgenic mouse. The lung of a F1 mouse (6 months old from line #71) was collected for histological analysis. The section shown here is a partial lobe of the lung (406 magnification). (c) Morphological features of lung tumors. The enlarged tumor was from an area pointed by an arrow shown in (b) (4006 magnification). All tissues were stained with hematoxylin-eosin solution and photographed

and cytoplasmic granular patterns. By analysing the immunoreactivity, it seems that the cancerous cells at the edges of the tumor displayed more intensive staining than those in the center. No immunoreactivity was found in control tumor sections (Figure 2b), in which normal preimmunized rabbit IgG (from same rabbit) was used. These results suggest that the immunoreactivity in tumors is specific to RON. To determine if overexpressed RON is constitutively active in vivo, the tyrosine phosphorylation assay was performed using lysates prepared from more than 20 individual lung tumors in Western blot analysis. Results showed that RON is autophosphorylated in tumors as determined in Western blot analysis (data not shown). Similar results were also found in cell lines expressing high levels of RON in the absence of MSP, suggesting that overexpression leads to the receptor activation. However, MSP at relatively low levels (7 – 19+5 ng/50 mg proteins) was detected in Western blot analysis in 13 of 13 lysates prepared from tumor samples, indicating that RON activation could also be induced by MSP in vivo. In addition, we determined if activated RON is caused by point mutations. The cDNA fragments corresponding to the kinase domain of the RON transgene from more than 20 lung tumor samples was sequenced. No activating mutations were detected (data not shown), indicating that RON activation is not caused by generating oncogenic forms of RON. These data demonstrated that RON is highly expressed and constitutively active in lung tumors in SPCp – RON transgenic mice. Overexpression and constitutive activation of wild-type RON could be the driving force leading to the tumor initiation and progression. To further characterize tumors, experiments aimed at identifying the cell-type origin of lung tumors were performed. Our attention was focused on type II and Clara cells because it has been shown that the SPC promoter functions specifically in type II and Clara cells (Glasser et al., 1990; Wikenheiser et al., 1992). The specific cellular marker for type II cells is SPC, which is produced only by type II cells (Mason et al., Oncogene

Figure 2 Detection of human RON expression in lung tumors from SPCp-RON transgenic mice. Immunohistochemical staining was conducted as described (Santoro et al., 1998). Lung sections from a F1 mice (lines #71) were treated with 15% acetic acid to block endogenous alkaline phosphatase and blocked with 1% BSA in Tris buffer (10 mm Tris, pH 7.4, with 150 mM NaCl, 1% BSA, and 0.1 Tween 20). Rabbit IgG antibody to human RON (Chen et al., 2000) was used as primary antibody. Goat anti-rabbit IgG conjugated with alkaline phosphatase was used as detecting antibody. The substrate for the reaction was Fast Red. Positive controls were MDCK cells expressing RON (data not shown) (Wang et al., 1994). RON was stained positively in tumors (a). Lung sections incubated with preimmunized rabbit IgG served as the negative control (b). The results were examined under a microscope and photographed. Similar results were obtained from lung tumor section derived from other lines of transgenic mice (data not shown)

2000). In contrast, nonciliated bronchiolar Clara cells express CC10 (Singh and Katyal, 1998), which serves as a distinct cellular marker. Thus, identifying type II

Lung tumors in RON transgenic mice Y-Q Chen et al

6385

or Clara cell markers in lung tumors could help us to determine the origin of cancerous cells. IHCS was performed using rabbit anti-proSPC serum and goat anti-CC10 serum (Singh and Katyal, 1998). The results are shown (Figure 3). Positive staining of the bronchial epithelial cells with the anti-CC10 antibody was observed (Figure 3a). No tumor cells stained positively with CC10. The results of detecting SPC are shown (Figure 3b). Positive staining of tumor cells with antiSPC antibody was observed. As a control, type II cells from littermates also expressed SPC (data not shown). Thus, lung tumors do not express the Clara cell marker CC10, but display type II cell phenotypes. It is very possible that lung tumors in the transgenic mice originated from type II pneumocytes. The role of RON in the initiation and progression of epithelial tumors is largely unknown. Previous studies have shown that forced expression of wtRON in rodent fibroblasts does not cause cell transformation, instead it induces motile-invasive phenotypes (Santoro et al., 1996). However, RON has oncogenic potentials (Santoro et al., 1998). It was shown that RON is overexpressed in the majority of primary breast carcinoma samples and in colon cancer cell lines (Maggiora et al., 1998; Chen et al., 2000). Also, the oncogenic forms of RON can be created by point mutations in the kinase domain of RON, which causes cell-transformation and tumor growth in vivo (Santoro et al., 1998). These data clearly suggest that under

Figure 3 Expression of type II cell marker SPC by lung tumors. Immunohistochemical staining was performed as described in Figure 2. (a) CC10 is expressed in bronchiole epithelial cells but not in tumors. The lung tumor sections from F1 mice were incubated with goat IgG against rabbit CC10 (cross-reacts with mouse CC10). Normal goat IgG was used as negative controls (data not shown). (b) SPC is detected in lung tumors from SPCpRON transgenic mice. The rabbit IgG antibody to rat proSPC was used as primary antibody. Lung sections same as those in (a) were used in this experiment. Normal rabbit IgG was used as negative controls (data not shown)

certain conditions, altered RON expression could initiate transforming/oncogenic programs. The data presented in this study show that overexpression of RON in distal lung epithelial cells results in lung tumor formation and progression in transgenic mice. However, it is acknowledged that significant difference exists between the cell transformation assay and the transgenic overexpression study. As shown in Table 1 and Figure 2, RON is highly expressed and constitutively activated in lung tumors in vivo. The activation is probably caused by abnormal receptor accumulation or through interaction with MSP. Activated RON has the ability to activate many signaling pathways including Ras (Li et al., 1995), PI-3 kinase (Wang et al., 1996), MAP kinase (Santoro et al., 1998), and JNK (Chen et al., 2000), all of which are critical in regulating cell growth, transformation, and survival. Thus, it is possible that constitutive activation of RON facilitates the uncontrolled growth of distal lung epithelial cells towards formation of lung adenomas. There are several features in lung tumors in SPCp – RON transgenic mice that differ from those induced by oncogenic proteins as described in literature (Wikenheiser et al., 1992; Lebel et al., 1995; Lavigueur et al., 1989; Kerkhoff et al., 2000; Clark et al., 2001). (1) RON is a receptor tyrosine kinase that is normally expressed in the lungs (Ronsin et al., 1993). Unlike SV40 and mutated H-Ras, RON is not an oncogenic protein nor does it have cell-transforming activity in vitro (Santoro et al., 1996). Thus, the mechanisms by which RON expression results in tumor formation must be different from those induced by oncogenic proteins such as SV40 (Wikenheiser et al., 1992). (2) Tumors in SPCp-RON mice have a unique cell morphology and growth pattern. They originate from alveolar, only express the type II cell marker SPC, and grow along the alveolar septa and project into the alveolar spaces. These features are different from those in multifocal lung tumors in SV40T transgenic mice (Sandmoeller et al., 1995). SV40T-induced tumors were adenocarcinomas with a mixed growth pattern: lepidic, papillary, and solid (Wikenheiser et al., 1992). Tumors also display phenotypes of several subtypes of adenocarcinomas with mixed cellular markers including CC10 and SPC, suggesting heterogeneity of tumors from bronchiolar and alveolar origins (Wikenheiser et al., 1992). (3) The progression of lung tumors in SPCp – RON mice is a slow process. Tumors were initiated in the lung as a single mass around 2 to 3 months after mice were born. Multiple nodules were observed only at 6 – 10 months and lasted 12 – 24 months. Mice that died due to respiratory failure usually lived for 2 years. This progression pattern differs from other transgenic lung tumor models. In those models, tumors developed rapidly within months and mice died around 4 – 6 months of age (Wikenheiser et al., 1992; Lavigueur et al., 1989; Lebel et al., 1995; Sandmoeller et al., 1995), mostly before transgenic descendants could be obtained. Thus, lung tumors in RON transgenic mice progress on a relatively natural course. Oncogene

Lung tumors in RON transgenic mice Y-Q Chen et al

6386

At present, the mechanisms by which RON expression results in lung tumors are unknown. Nevertheless, it is worthwhile to speculate the potential mechanisms underlying this unique tumor model. Since wtRON does not have cell-transforming activities in vitro (Santoro et al., 1996), it is unlikely that the tumors are caused directly by overexpression of RON in vivo. A second hit with additional genomic or cellular changes might be required for the complete initiation of lung tumors in SPCp – RON transgenic mice. Our recent studies have found that genomic instability occurs in individual lung tumors. Furthermore, lung tumors were not found in newborn mice but only in 2 – 3 months old mice, suggesting that a 60 – 90 day period is required for tumors to develop in the lung. This period may provide opportunities for alterations in some oncogenes or tumor suppressor genes in RON expressing type II cells. Also, lung tumors in young mice (3 – 8 months old) are usually adenomas with some atypical cells showing early progression towards adenocarcinomas. In contrast, tumors in relatively old mice (10 – 24 months) show increased malignant

behavior, indicating that additional changes are needed for adenomas to develop into adenocarcinomas. Mutations or deletions in oncogenes or tumor suppressor genes could facilitate this progression. It is often reported that K-Ras or p53 gene is frequently mutated in mouse lung tumors, induced either by chemical carcinogens or by expression of oncogenes (Malkinson, 1998). Thus, it will be important in the future to study the molecular and cellular mechanisms involved in the initiation and progression of lung tumors in SPCp-RON transgenic mice. Acknowledgments We thank Drs JA Whitsett (University of Cincinnati, Cincinnati, OH, USA) for the 3.7SPC/SV40 vector, G Singh (Department of Veterans Affairs Medical Center, Pittsburgh, PA, USA) for antibodies to CC10 and proSPC, and W Gunning (Medical College of Ohio, Toledo, OH, USA) for histological examination. We are indebted to Dr Y-S Ho for his kind help in the generation of the transgenic mice. This work was supported by NIH Grants RO1 AI43516 and CA91980 to M-H Wang.

References Angeloni D, Danilkovitch-Miagkova A, Ivanov SV, Breathnach R, Johnson BE, Leonard EJ and Lerman MI. (2000). Genes Chrom. Cancer, 29, 147 – 156. Chen Y-Q, Zhou Y-Q, Angeloni D, Kurtz AL, Qiang X-Z and Wang M-H. (2000). Exp. Cell Res., 261, 229 – 238. Clark JC, Tichelaar JW, Wert SE, Itoh N, Perl A-KT, Stahlman MT and Whitsett JA. (2001). Am. J. Physiol. Lung Cell Mol. Physiol., 280, L705 – L715. Gaudino G, Avantaggiato V, Follenzi A, Acampora D, Simeone A and Comoglio PM. (1995). Oncogene, 11, 2627 – 2637. Gaudino G, Follenzi A, Naldini L, Collesi C, Santoro M, Gallo KA, Godowski PJ and Comoglio PM. (1994). EMBO J., 13, 3524 – 3532. Glasser SW, Korfhagen TR, Bruno MD, Dey C and Whitsett JA. (1990). J. Biol. Chem., 265, 21986 – 21991. Han S, Stuart LA and Degen SJF. (1991). Biochemistry, 30, 9768 – 9780. Ho Y-S, Vincent R, Dey MS, Slot JW and Crapo JD. (1998). Am. J. Respir. Cell. Mol. Biol., 18, 538 – 547. Hogan BF, Constantiti F and Lacy E. (1996). Manipulating the Mouse Embryo. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. Iwama A, Okano K, Sudo T, Matsuda Y and Suda T. (1994). Blood, 83, 3160 – 3169. Kerkhoff E, Fedorov LM, Siefken R, Walter AO, Papadopoulos T and Rapp UR. (2000). Cell Growth Differ., 11, 185 – 190. Lavigueur A, Maltby V, Mock D, Rossant J, Pawson T and Bernstein A. (1989). Mol. Cell Biol., 9, 3982 – 3991. Lebel M, Webster M, Muller W, Royal A, Gauthier J and Mes-Masson A-M. (1995). Cell Growth Differ., 6, 1591 – 1600. Li B, Wang M-H, Kung H-F, Ronsin C, Breathnach R, Leonard EJ and Kamata T. (1995). Biochem. Biophys. Res. Commun., 216, 110 – 118. Maggiora P, Marchio S, Stella MC, Giai M, Belfiore A, De Bortoli M, Di Renzo MF, Costantino A, Sismondi P and Comoglio PM. (1998). Oncogene, 16, 2927 – 2933. Oncogene

Malkinson AM. (1998). Exp. Lung Res., 24, 541 – 555. Mason RJ, Kalina M, Nielsen LD, Malkinson AM and Shannon JM. (2000). Am. J. Pathol., 156, 175 – 182. Muraoka RS, Sun WS, Colbert MC, Waltz SE, Witte DP, Degen JL and Degen SJF. (1999). J. Clin. Invest., 103, 1277 – 1285. Ronsin C, Muscatelli F, Mattei MG and Breathnach R. (1993). Oncogene, 8, 1195 – 1202. Sakamoto O, Iwama A, Amitani R, Takehara T, Yamaguchi N, Yamamoto T, Masuyama K, Yamanaka T, Ando M and Suda T. (1997). J. Clin. Invest., 99, 701 – 709. Sandmoeller A, Halter R, Suske G and Beato M. (1995). Cell Growth Differ., 6, 97 – 103. Santoro MM, Collesi C, Grisendi S, Gaudino G and Comoglio PM. (1996). Mol. Cell. Biol., 16, 7072 – 7083. Santoro MM, Penengo L, Minetto M, Orecchia S, Cilli M and Gadino G. (1998). Oncogene, 17, 741 – 749. Singh G and Katyal S. (1998). Am. J. Respir. Cell Mol., 17, 141 – 143. Skeel A, Yoshimura T, Showalter SD, Tanaka S, Appella E and Leonard EJ. (1991). J. Exp. Med., 173, 1227 – 1234. Wang M-H, Ronsin C, Gesnel MC, Coupeym L, Skeel A, Leonard EJ and Breathnach R. (1994). Science, 266, 117 – 119. Wang M-H, Montero-Julian FA and Leonard EJ. (1996). Oncogene, 13, 2167 – 2175. Wikenheiser KA, Clark JC, Linnoila I, Stahlman MT and Whitsett JA. (1992). Cancer Res., 52, 5342 – 5352. Willett C, Wang M-H, Emanuel RL, Graham SA, Smith DI, Shridhar V, Sugarbaker DJ and Sunday M. (1998). Am. J. Respir. Cell Mol. Biol., 18, 489 – 496. Willett CG, Smith DI, Shridhar V, Wang M-H, Emanuel R, Patidar K, Graham SA, Zhang F, Hatch V, Sugarbaker JD and Sunday ME. (1997). J. Clin. Invest., 99, 2979 – 2991.