当前位置: 首页 > 医学版 > 期刊论文 > 基础医学 > 免疫学杂志 > 2005年 > 第11期 > 正文
编号:11256763
Overexpression of Stat3C in Pulmonary Epithelium Protects against Hyperoxic Lung Injury1
     Abstract

    Acute lung injury is a side effect of therapy with a high concentration of inspired oxygen in patients. The molecular mechanism underlining this effect is poorly understood. In this study, we report that overexpression of Stat3C, a constitutive active form of STAT3, in respiratory epithelial cells of a doxycycline-controlled double-transgenic mouse system protects lung from inflammation and injury caused by hyperoxia. In this mouse line, >50% of transgenic mice survived exposure to 95% oxygen at day 7, compared with 0% survival of wild-type mice. Overexpression of STAT3C delays acute capillary leakage and neutrophil infiltration into the alveolar region. This protection is mediated at least partially through inhibition of hyperoxia-induced synthesis and release of matrix metalloproteinase (MMP)-9 and MMP-12 by neutrophils and alveolar resident cells. In some MMP-9–/– mice, prolonged survival was observed under hyperoxic condition. The finding supports a concept that activation of the Stat3 pathway plays a role to prevent hyperoxia-induced inflammation and injury in the lung.

    Introduction

    Signal transducer and activator of transcription 3 is originally identified as the acute-phase response factor (1, 2). Stat3 is mainly activated by autocrine/paracrine growth factors and IL-6 family cytokines that share the common gp130 receptor subunit (3, 4). Upon activation by cytokines or growth factors, phosphorylation at Y705 by JAKs causes Stat3 dimerization to form heterodimer with Stat1, or to form homodimer with itself. Stat3 dimers translocate into the nucleus and activate downstream target genes. A Stat3C form has been designed to mimic the action of phosphorylated Stat3 (5). In this molecule, substitution of two cysteine residues within the C-terminal loop of the Src homology 2 domain of Stat3 produces a molecule that dimerizes spontaneously, binds to DNA, and activates gene transcription. The Stat3 signaling stimulates cell proliferation and prevents apoptosis (5). Overstimulation of the Stat3 pathway causes cellular transformation in mice (5). In contrast to other Stat-deficient mice that are born alive, targeted disruption of the mouse Stat3 gene leads to early embryonic lethality (6). Therefore, Stat3 is required for normal embryonic development and tissue formation. In the lung, Stat3 is expressed in alveolar type II epithelial cells and Clara cells along with IL-6Rs and JAKs (7, 8). Treatment of IL-6 or LPS activates Stat3 phosphorylation at Y705 in these cells in vivo (8, 9). Stat3 is required for maintaining the alveolar epithelial structure and function. Overexpression of a dominant-negative Stat3 to subvert the endogenous Stat3 activity in respiratory epithelial cells causes alveolar destruction (8). This pathophysiological phenotype is mediated partially through Stat3 regulation of surfactant protein B (SP-B)3 homeostasis in the lung. In the absence of SP-B, neonatal lungs are collapsed during respiratory cycles that cause death after birth (10, 11). Stat3 stimulates SP-B gene expression in respiratory epithelial cells in a synergistic manner with retinoid acid receptors (7, 8). In an animal system, in which the Stat3 allele is selectively deleted in respiratory epithelial cells, surfactant proteins and lipids are decreased or absent in alveolar lavage material during hyperoxia. Pulmonary surfactant reduces surface tension forces and protects lung from collapse during respiratory cycles. Addition of SP-B to this system can improve animal survival during oxygen injury (12).

    Supplemental oxygen is commonly administered to patients with acute and chronic cardiopulmonary disorders to augment tissue oxygen delivery and enhance the alveolar and arterial oxygen levels. Unfortunately, prolonged administration of inspired oxygen at >50% concentration leads to a variety of forms of lung tissue damage, including acute lung injury and adult respiratory distress syndrome (13, 14). Animals exposed to hyperoxic conditions have been used as models to study acute lung injury and adult respiratory distress syndrome (13). There are three stages of morphologic changes in mouse models after exposure to a lethal dose of hyperoxia (e.g., 95–100% O2). During the first 24 to 72 h of exposure, most animals show no significant morphologic changes in the alveolar septa. This phase is soon followed by an inflammatory phase, which is characterized by an infiltration and accumulation of inflammatory cells (such as neutrophils) in the lung and the release of inflammatory mediators. This leads to remodeling and destruction of alveolar parenchyma that directly cause animal death (a destructive phase). The molecular mechanism involved in this whole process is poorly understood.

    In the mature lung, the interalveolar walls are supplied with capillaries. Epithelial cells and endothelial cells form a thin alveolar wall for easy penetration and exchange of oxygen and carbon dioxide. Extracellular matrix (ECM) is the major component in the endothelial-epithelial interstitial structure that is comprised of multiple structural proteins (such as collagens, elastin, gelatin, etc.). ECM can be modified by matrix metalloproteinases (MMPs) that are zinc-dependent matrix-degrading proteinases. They degrade almost every component of the ECM and are required for normal lung development and remodeling after lung injury. However, exuberant or aberrant expression of MMPs can cause tissue damage (15). It is conceivable that MMPs play a major role in the alveolar destruction during hyperoxia. Under normal physiological conditions, the activities of MMPs are regulated at the level of transcription, posttranslational processing, and inhibition by endogenous inhibitors. Tissue inhibitors of metalloproteinases (TIMPs) are specific inhibitors of MMPs that participate in controlling the local activities of MMPs in tissues. Aberrant expression of TIMPs can also lead to tissue damage.

    Because inhibition of the endogenous Stat3 activity causes alveolar destruction and accelerates animal death during hyperoxia (8), it suggests that activation of the Stat3 pathway is required to protect against oxygen injury caused by hyperoxia. To test this hypothesis, an animal model has been established, in which a constitutively active form Stat3C is overexpressed in respiratory epithelial cells under the control of a tissue-specific promoter by doxycycline treatment. Using this animal model system, we discovered that activation of the Stat3 pathway in respiratory epithelial cells protects against hyperoxia-induced capillary hemorrhage and neutrophil influx into the lung. In this animal model system, expression and enzymatic activities of MMP-9 and MMP-12 were highly induced during hyperoxia. Interestingly, these pathogenic increases were suppressed by Stat3C overexpression. As a result, the survival rate of transgenic mice was significantly improved during hyperoxia. The observation identifies a new mechanism by which the Stat3 signaling protects against the hyperoxia-induced lung inflammation and injury.

    Materials and Methods

    Mice were anesthetized and lungs were inflation-fixed with 4% paraformaldehyde in PBS overnight at 4°C. Lungs were washed with PBS and dehydrated through a series of ethanol, followed by paraffin embedding. Five-micrometer tissue sections were loaded onto slides for H&E staining and immunohistochemistry. In immunohistochemistry, MMP-9 Ab (1:500) (18) and MMP-12 Ab (1:200; Santa Cruz Biotechnology) were used. The signals were detected by biotinylated secondary Abs and Vectastain Elite ABC kit (Vector Laboratories). For morphometric analyses, images were taken from various lung lobes and transferred by video camera to a computer screen using MetaMorph imaging software (Universal Imaging). A computer-generated, 121-point lattice grid was superimposed on each field, and numbers of MMP-9-stained cells were determined by a point counting method. The statistical significance was assessed by ANOVA.

    For immunofluorescence staining, Ly6G Ab (BD Biosciences) and MMP-9 Ab were used as primary Abs. A Cy2-conjugated donkey anti-rabbit IgG and a Cy3-conjugated donkey anti-rat IgG (Jackson ImmunoResearch) at a dilution of 1/200 were used as the secondary Abs. Slides were examined under a Nikon fluorescent microscope and analyzed for red (Cy3) and green (Cy2) fluorescence.

    Results

    A doxycycline-controlled double-transgenic mouse model was generated to specifically direct Stat3C expression in respiratory epithelial cells. In this system, a previously made CCSP-rtTA transgenic mouse line (16) was crossbred with a newly generated teto-CMV-Stat3C transgenic mouse line (Fig. 1). A Flag sequence was added at the C terminus of the Stat3C cDNA. Double-transgenic mice have been successfully obtained as detected by PCR using specific primers and mouse tail DNAs.

    Stat3C overexpression enhances animal survival under hyperoxic condition

    To determine how constitutive activation of the Stat3 signaling pathway by Stat3C overexpression affects animal survival under oxygen toxicity, doxycycline-treated or -untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic and WT mice were exposed to 95% oxygen. Under this condition, the first mortality of doxycycline-treated or -untreated WT mice began at day 5. At day 7, all of the WT mice had succumbed (Fig. 3). Because both doxycycline-treated and -untreated WT animals showed identical survival curve, it seems that doxycycline treatment has no effect on animal survival during hyperoxia. The doxycycline-untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice showed a similar survival curve like WT animals (data not shown). In comparison, induction of Stat3C overexpression by doxycycline treatment in CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice significantly delayed mortality under the same condition. The onset of death for doxycycline-treated double-transgenic mice began at day 6. At day 7, >50% of transgenic mice were still surviving 95% oxygen exposure, compared with 0% survival in WT mice. Some doxycycline-treated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic animals survived until day 9. Therefore, overexpression of Stat3C in respiratory epithelial cells enhances survival following exposure to hyperoxia.

    Stat3C overexpression protects against pulmonary hemorrhage under hyperoxic condition

    To further address the morphological changes caused by hyperoxia in the lung, histopathologic assessment was performed. Doxycycline-treated or -untreated WT and CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic animals were exposed to 95% oxygen for 4.5 days (right before the onset of mortality in WT animals). Lungs were inflated and paraffin embedded. After staining with H&E, tissue sections from WT and CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic animals were compared. In WT animals, alveolar hemorrhage was apparent after 95% oxygen exposure regardless of doxycycline treatment, indicating increased capillary permeability and damage of the endothelial-epithelial interstitial structure along the alveolar wall (Fig. 4). The doxycycline-untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice showed a similar result (data not shown). However, induction of Stat3C overexpression after doxycycline treatment of CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice showed no obvious hemorrhage under the same hyperoxic condition (Fig. 4).

    Stat3C overexpression reduces MMP-9 and MMP-12 mRNA levels under hyperoxic condition in the lung

    Pulmonary hemorrhage during hyperoxia indicates alveolar structure damage, which may be caused by elevation of MMPs in the lung. To test this assumption, RT-PCR assay of MMP-9 and MMP-12 mRNAs was performed. Doxycycline-treated or -untreated WT or CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic animals were exposed to 95% oxygen for 4.5 days. Lungs were isolated and total RNAs were purified. Regardless of doxycycline treatment, MMP-9 and MMP-12 mRNA expression levels were significantly increased after 95% oxygen exposure in WT mice (Fig. 5). In doxycycline-untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice, MMP-9 and MMP-12 mRNA expression levels were significantly increased after 95% oxygen exposure, which were similar to WT mice (Fig. 5). However, induction of Stat3C overexpression by doxycycline treatment of CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice significantly inhibited elevation of MMP-9 and MMP-12 mRNA levels in the lung (Fig. 5).

    Stat3C overexpression protects against neutrophil infiltration under hyperoxic condition in the lung

    Neutrophil infiltration is a manifestation of many acute inflammatory responses in the lung (20). Proteinases including MMP-9 (gelatinase B) are synthesized and stored in granules of neutrophils, and released upon inflammatory stimulation (21). Using polyclonal Ab against MMP-9 as a marker described previously (18), MMP-9 influx into the lung was observed in both doxycycline-treated or -untreated WT animals after 4.5 days of 95% oxygen exposure. The process includes adhesion of MMP-9-positive cells on the endothelial surface of blood vessels, infiltration into the perivascular space region and the alveolar interstitial region (Fig. 6, A and B (pointed by arrows)). The staining of doxycycline-untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice was similar to that of WT mice (data not shown). In comparison, MMP-9-positive cell influx was significantly reduced in doxycycline-treated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice, in which Stat3C was overexpressed.

    MMP-9 influx is associated with neutrophil infiltration

    To confirm that MMP-9-positive cells are indeed neutrophils, a colocalization of double immunofluorescent staining was performed using Abs against MMP-9 and Ly6G (a neutrophil-specific marker). In oxygen-untreated WT mice, no obvious MMP-9 or Ly6G staining was observed (Fig. 7, upper three panels). Oxygen treatment significantly increased MMP-9- and Ly6G-positive staining cells in the lung (Fig. 7, middle three panels). The overlay of MMP-9- and Ly6G-positive cells was 100% identical in the oxygen-treated WT lung (Fig. 7, middle right panel). This is a clear indication that the MMP-9-positive cells in oxygen-treated mice are neutrophils. In contrast, in doxycycline-treated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice, both MMP-9- and Ly6G-positive cells were significantly reduced (Fig. 7, bottom panels).

    Stat3C overexpression decreases hyperoxia-induced MMP-9 enzymatic activity in the lung

    To determine whether neutrophil and MMP-9 influx results in elevation of MMP-9 enzymatic activity in the alveolar lumen, BALF were collected from doxycycline-treated WT and CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice, oxygen exposed or unexposed, for zymography analysis. As shown in Fig. 8, the MMP-9 activity was dramatically increased in WT animals after oxygen exposure. Doxycycline-untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice showed a result similar to that of WT mice (data not shown). In comparison, induction of Stat3C overexpression blocked increase of the MMP-9 activity in doxycycline-treated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice. Therefore, elevation of the MMP-9 activity is partially responsible for damage of the endothelial-epithelial interstitial structure and capillary leakage in alveoli. This study suggests that activation of the Stat3 pathway blocks the MMP-9-damaging effect by delaying neutrophil influx into the alveolar region.

    Stat3C overexpression decreases hyperoxia-induced MMP-12 synthesis in resident cells around the alveolar wall

    MMP-12 is another important proteinase that degrades elastin and type IV collagen (15, 22). Previous studies showed that MMP-12 is primarily secreted by macrophages and plays a role in the pathogenesis of chronic lung injury, especially in smoking-related chronic obstructive pulmonary disease and emphysema (23, 24). Using polyclonal Ab against MMP-12, MMP-12 expression was significantly induced in the resident cells along the alveolar wall in doxycycline-treated or -untreated WT animals after 4.5 days of 95% oxygen exposure on immunohistochemical staining (Fig. 10). The doxycycline-untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice showed a similar result (data not shown). In contrast, Stat3C overexpression blocked MMP-12 expression in doxycycline-treated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic animals in the same condition, indicating that activation of the Stat3 pathway suppresses the regional MMP de novo synthesis in resident cells along the alveolar wall.

    Stat3C overexpression has no protective effect on TIMP1 mRNA down-regulation under hyperoxic condition in the lung

    TIMPs are specific inhibitors for MMPs in tissues. One possible mechanism to explain alveolar structure damage and pulmonary hemorrhage during hyperoxia is down-regulation of TIMPs. To test this possibility, doxycycline-treated or -untreated CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic animals were exposed to 95% oxygen for 4.5 days. Total RNAs were isolated from whole lungs and evaluated by RT-PCR. Among four TIMPs, the TIMP1 mRNA expression level was significantly decreased after 95% oxygen exposure (Fig. 11). The TIMP3 mRNA expression level remained the same. There was no detectable TIMP2 or TIMP4 expression in the lung regardless of oxygen treatment. The WT animals showed a similar result (data not shown). This result suggests that down-regulation of TIMP1 plays a role in alveolar structure damage during hyperoxia. However, unlike MMP-9 or MMP-12, induction of Stat3C overexpression by doxycycline treatment of CCSP-rtTA/(teto)7-CMV-Stat3C double-transgenic mice showed no protection against down-regulation of TIMP1 mRNA expression in the lung (Fig. 11).

    Discussion

    Oxygen-induced lung injury is a complex process, which is affected by multiple factors. The mechanism underlying this process is not very clear. Because pulmonary hemorrhage is a major histopathologic finding in mice treated with 95% oxygen (Fig. 4), the interstitial structure of endothelial-epithelial cells must be damaged during hyperoxia. As reported previously, hyperoxia decreases pulmonary surfactant (12). This does not directly lead to ECM degradation and pulmonary hemorrhage. Other mechanisms must be involved in the process.

    MMPs are the major enzymes that degrade ECM proteins in the lung. It is reasonable to speculate that the damage of the alveolar interstitial structure results from the elevated levels of MMPs in oxygen-treated mice. Indeed, the RT-PCR and zymography studies showed dramatic elevation of the MMP-9 and MMP-12 levels in the WT lung before the onset of animal mortality following hyperoxia exposure (Fig. 5). MMP-9 can be synthesized by multiple cell types in the lung (21). The increase of the MMP-9 activity may come from either the migrating circulatory cells (e.g., neutrophils), or from alveolar resident cells. To clarify this issue, immunohistochemical and immunofluorescent staining was performed. The results indicated that MMP-9 staining is tightly associated with invading neutrophils from the circulatory system (Figs. 6 and 7). This occurs in both the alveolar region and the blood vessel region in the lung. It seems that hyperoxia triggers neutrophil adhesion to endothelial cells and infiltration into the lung. The molecular mechanism for this phenomenon remains to be elucidated. No MMP-9 expression was detected in alveolar resident cells. In comparison, newly synthesized MMP-12 was detected in resident cells along the alveolar walls after hyperoxia exposure (Fig. 10). MMP-12 is generally regarded as macrophage elastase and is secreted by macrophages. There was no obvious macrophage increase in the alveolar region during hyperoxia. In addition to MMP elevation, expression of an MMP inhibitor TIMP1 was decreased after hyperoxia exposure (Fig. 11). Taken together, hyperoxia enhances MMP activities in alveoli either by inducing MMP expression or by inhibiting TIMP expression. As a consequence, these enzymes directly cause ECM destruction, leading to capillary permeability and subsequent pulmonary hemorrhage.

    Because oxygen injury is initiated by inflammation, molecules that play important roles in mediating inflammatory responses may participate in the process of oxygen injury. Transgenic mouse model systems have been proven to be very useful in identifying signaling molecules and transcription factors in various lung disease processes. In this study, a doxycycline-controlled Stat3C double-transgenic mouse system has been created. In this system, the Stat3 signaling pathway is constitutively activated by overexpression of Stat3C. The studies showed that activation of the Stat3 signaling pathway in respiratory epithelial cells exhibits the protective effect against oxygen injury that is associated with increased capillary permeability, neutrophil invasion, and MMP synthesis in the lung. However, this pathway does not protect against down-regulation of TIMP1 expression caused by hyperoxia.

    Based on the observation made in this report, a model has been proposed to explain the molecular mechanism by which Stat3C protects against lung oxygen injury. High oxygen concentration induces expression of MMPs (e.g., MMP-12) in lung epithelial cells, endothelial cells, and bronchoalveolar macrophages. These gene products cause ECM degradation. In addition, high oxygen concentration induces expression of signaling molecules that bind to the cognate plasma membrane receptors and results in the increased expression of adhesion molecules in endothelial cells to facilitate migration and adhesion of neutrophils to the vascular endothelium. Attached neutrophils migrate through the capillary wall into the lung. The activated neutrophils in the lung release proteinases (e.g., MMP-9) to exacerbate capillary permeability and lung injury. When Stat3C is overexpressed in alveolar type II epithelial cells, it activates a set of genes. Through paracrine and autocrine mechanisms, these gene products are released from alveolar type II epithelial cells to inhibit influx of neutrophils and production of MMP-9 and MMP-12 in the lung. Therefore, alveolar type II epithelial cells play a critical role in damaging and repairing of lung oxygen injury.

    Other mechanisms may also be involved in the protection against oxygen-induced lung injury. SP-B is a critical component for the surfactant structure and is required for epithelial cell remodeling following the oxygen injury (25). In CCSP-rtTA/(teto)7-CMV-Stat3C transgenic mice, overexpression of Stat3C in the lung promoted synthesis of SP-B that protects and maintains the surfactant integrity (Fig. 2). Elevated production of SP-B in doxycycline-treated CCSP-rtTA/(teto)7-CMV-Stat3C transgenic mice helps to resist alveolar structure damage during hyperoxia. This is in agreement with a previous report that expression of Stat3 in respiratory epithelial cells plays a critical role in maintenance of surfactant protein and lipid homeostasis during oxygen injury (12). Addition of SP-B to this system through the intratracheal route can improve survival during oxygen injury (12). Taken together, expression of Stat3 protects against oxygen-induced lung injury through at least two mechanisms. First, it decreases MMP production and unwanted neutrophil influx into the lung. Second, it promotes surfactant protein and lipid synthesis to maintain the integrity of the alveolar structure and to prevent alveoli from collapse during hyperoxia injury.

    Footnotes

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

    1 This work was supported by National Heart, Lung, and Blood Institute/National Institutes of Health (NHLBI/NIH) Grant 061803 (to C.Y. and H.D.), NHLBI/NIH Grant 067862 (to C.Y. and H.D.), March of Dimes Grant FY02-206 (to C.Y.), NHLBI/NIH Grant 47328 (to R.M.S.), NHLBI/NIH 29594 (to R.M.S.), and the Alan A. and Edith L. Wolff Charitable Trust (to R.M.S.).

    2 Address correspondence and reprint requests to Dr. Cong Yan, Cincinnati Children’s Hospital Medical Center, Division of Pulmonary Biology, The Children’s Hospital Research Foundation, 3333 Burnet Avenue, Cincinnati, OH 45229-3039. E-mail address: Cong.Yan{at}cchmc.org

    3 Abbreviations used in this paper: SP-B, surfactant protein B; ECM, extracellular matrix; MMP, matrix metalloproteinase; TIMP, tissue inhibitor of metalloproteinase; TRE, Tet-responsive element; rtTA, reverse tetracycline-responsive transactivator; WT, wild type; BALF, bronchoalveolar lavage fluid.

    Received for publication October 6, 2004. Accepted for publication March 14, 2005.

    References

    Akira, S., Y. Nishio, M. Inoue, X. J. Wang, S. Wei, T. Matsusaka, K. Yoshida, T. Sudo, M. Naruto, T. Kishimoto. 1994. Molecular cloning of APRF, a novel IFN-stimulated gene factor 3 p91-related transcription factor involved in the gp130-mediated signaling pathway. Cell 77: 63-71.

    Wegenka, U. M., C. Lutticken, J. Buschmann, J. Yuan, F. Lottspeich, W. Muller-Esterl, C. Schindler, E. Roeb, P. C. Heinrich, F. Horn. 1994. The interleukin-6-activated acute-phase response factor is antigenically and functionally related to members of the signal transducer and activator of transcription (STAT) family. Mol. Cell. Biol. 14: 3186-3196.

    Leonard, W. J., J. J. O’Shea. 1998. Jaks and STATs: biological implications. Annu. Rev. Immunol. 16: 293-322.

    Taga, T., T. Kishimoto. 1997. Gp130 and the interleukin-6 family of cytokines. Annu. Rev. Immunol. 15: 797-819.

    Bromberg, J. F., M. H. Wrzeszczynska, G. Devgan, Y. Zhao, R. G. Pestell, C. Albanese, J. E. Darnell, Jr. 1999. Stat3 as an oncogene. Cell 98: 295-303.

    Takeda, K., K. Noguchi, W. Shi, T. Tanaka, M. Matsumoto, N. Yoshida, T. Kishimoto, S. Akira. 1997. Targeted disruption of the mouse Stat3 gene leads to early embryonic lethality. Proc. Natl. Acad. Sci. USA 94: 3801-3804.(Xuemei Lian*,, Yulin Qin*)