Available online at www.sciencedirect.com
Molecular Aspects of Medicine 28 (2007) 453–480 www.elsevier.com/locate/mam
Review
Tocopherol transfer protein deficiency modifies nuclear receptor transcriptional networks in lungs: Modulation by cigarette smoke in vivo K. Gohil *, S. Oommen, V.T. Vasu, H.H. Aung, C.E. Cross Pulmonary and Critical Care Medicine, Genome and Biomedical Sciences Facility, 451 East Health Sciences Drive, University of California, Davis, CA 95616, USA Received 2 February 2007; revised 12 February 2007; accepted 13 February 2007
Abstract Dietary factors and environmental pollutants initiate signaling cascades that converge on AhR:Nrf2:NF-jB transcription factor (TF) networks and, in turn, affect the health of the organism through its effects on the expression of numerous genes. Reactive oxygen metabolites (ROMs) have been hypothesized to be common mediators in these pathways. a-Tocopherol (AT) is a potent, lipophilic, scavenger of ROMs in vitro and has been hypothesized to be a major chain-breaking anti-oxidant in lipoproteins and biological membranes in vivo. The lung offers a vital organ to test the various postulated actions of AT in vivo. Lung AT concentrations can be manipulated by several methods that include dietary and genetic techniques. In this study we have used mice with severe AT deficiency inflicted at birth by the deletion of AT transfer protein (ATTP) which is abundantly expressed in the liver and regulates systemic concentrations of AT. Mice and humans deficient in ATTP are AT deficient. Female ATTP-deficient (ATTP-KO) mice and their congenic ATTP normal (WT) mice fed a diet containing 35 IU AT/kg diet were used to test our hypothesis. The mice (n = 5/group) were exposed to either air or cigarette smoke (CS, total suspended particles 60 mg/m3, 6 h/day), a source of ROM, for 3 or 10 days. Post-exposure lung tissue was dissected, RNA extracted from each lung and it was pooled group-wise and processed for GeneChip analysis (Affymetrix 430A 2.0). Differential analysis of the transcriptomes (16,000 mRNAs) identified CS sensitive genes that were modulated by lung AT-concentration. CS activated AhR driven genes such as cyp1b1 whose induction *
Corresponding author. Tel.: +1 530 754 6932; fax: +1 530 752 8632. E-mail address:
[email protected] (K. Gohil).
0098-2997/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.mam.2007.02.004
454
K. Gohil et al. / Molecular Aspects of Medicine 28 (2007) 453–480
was augmented in CS-exposed, AT-deficient lungs. However, CS-induced expression of some of the Nrf2 driven genes was not potentiated in the AT-deficient lungs. Largest clusters of CS-AT sensitive genes were lymphocyte and leukocyte specific genes. These gene-clusters included those encoding cytokines and immunoglobulins, which were repressed by CS and were modulated by lung AT concentrations. Our genome-wide analysis suggests reciprocal regulation of xenobiotic and immune response genes by CS and a modulatory role of lung AT concentration on the expression of these clusters of genes. These data suggest that in vivo network of AT, ATmetabolites and ATTP affects the transcription of genes driven by AhR, Nrf2 and NF-jB, transcription factor networks that transduce cellular metabolic signals and orchestrate adaptive responses of lungs to inhaled environmental pollutants. 2007 Elsevier Ltd. All rights reserved. Keywords: Cigarette smoke; Tobacco; d-Chip; Gene-networks; Immune response; Inflammation; Microarrays; a-Tocopherol transfer protein; Arntl; Nrf2; NF-jB; Vitamin E
Contents 1.
2.
3.
4.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Lessons from yeast: transcription of multiple genes is affected by nutrients and oxidant products of metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2. Genome-wide responses to AT in vivo . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.1. Transcriptional responses to dietary AT . . . . . . . . . . . . . . . . . . . . 1.2.2. Transcriptional responses to AT due to deletion of ATTP-gene . . . 1.2.3. Transcriptional responses of lungs to inhaled oxidants and xenobiotics of cigarette smoke (CS) . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. ATTP-KO and WT mice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. CS exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. RNA extraction and, GeneChip, quantitative real-time PCR (qRTPCR) and immunoblot analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Statistical data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. ATTP-KO mice do not express ATTP in liver or lungs . . . . . . . . . . . . . 3.2. Genome-wide responses of lungs to AT-deficiency and to CS . . . . . . . . . 3.2.1. Genome-wide responses of lungs to AT deficiency in the absence of ‘‘oxidative–xenobiotic’’ stress . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Genome-wide responses of lungs to oxidative and xenobiotic stress of CS in AT deficient lungs of ATTP-KO mice . . . . . . . . . . . . . . . 3.3. Induction of AhR driven genes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Modulation of Nrf2 driven genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Modulation of immune-inflammatory genes. . . . . . . . . . . . . . . . . . . . . . Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
455 455 456 456 457 457 458 458 459 459 459 459 459 460 460 463 466 469 470 473 474 474
K. Gohil et al. / Molecular Aspects of Medicine 28 (2007) 453–480
455
1. Introduction 1.1. Lessons from yeast: transcription of multiple genes is affected by nutrients and oxidant products of metabolism The invention and application of DNA and oligonucleotide microarray tools to quantify genome-wide expression of mRNAs have demonstrated that a single nutrient can generate complex transcriptional responses. The utility and power of these tools to define transcriptional networks operating in vivo is exemplified by studies in yeast (DeRisi et al., 1997; Lipshutz et al., 1995; Stengele et al., 2005). These studies offer a paradigm for investigations into defining the diversity of the actions of vitamin E in vivo. The yeast (Saccharomyces cerevisiae) genome is distributed over 16 chromosomes containing 5654 predicted protein coding genes (David et al., 2006). In yeast, repletion of glucose initiates a major transcriptional reprogramming in response to extracellular glucose (DeRisi et al., 1997; Kresnowati et al., 2006; Westergaard et al., 2006). The signaling cascade initiated by a change in extracellular glucose concentration converges on distinct domains of the genome and RNA polymerase complex containing 150 proteins to initiate the glucose sensitive transcriptional program. The changes in the transcription of most of the glucose sensitive genes can be accounted for by coordinated interactions of a few transcription factors (TFs), TF-network, on their specific targets on the genome. For example, glucose sensitive activation of 589 genes of the yeast genome could be accounted for by the activities of 12 TFs (Kresnowati et al., 2006). Although such comprehensive studies on the action of glucose in mammalian cells are lacking, genome-wide responses of b-pancreatic cells to glucose suggest that a large number of genes are modulated in response to glucose (Schuit et al., 2002); functional classification of the affected genes led the authors to suggest that glucose stimulated the conversion of mitochondrial metabolites into lipid intermediates. Transcriptional mechanisms that may co-ordinate glucose and lipid metabolism remain to be characterized and include recently discovered TFs such as carbohydrate responsive element binding protein (Dentin et al., 2006). It is likely that mammalian cells will have distinct cell specific, and yet to be defined, transcriptional responses to vital nutrients such as glucose and vitamins such as AT. Aerobic metabolism of glucose and other nutrients generates reactive oxygen metabolites (ROMs) such as hydrogen peroxide (H2O2) which is used as a molecular weapon to kill invading microbial pathogens (El-Benna et al., 2005), as a molecular signal that affects cell cycle and survival (Stone and Yang, 2006), and as a mediator of tissue damage (Bergamini et al., 2004; Frisard and Ravussin, 2006; Jaeschke, 2006; Xu and Touyz, 2006). Search for genome-wide transcriptional signatures of changes in H2O2 concentrations have identified a complex response of a genome to H2O2. The expression of 900 yeast genes are simultaneously affected by H2O2 and other oxidants (Gasch et al., 2000). Similarly, exposure of murine macrophages to H2O2 (Zhang et al., 2005) affected the expression of 113 genes that regulate cell survival, stress, and metabolism of glucose and lipids. The H2O2 stress also affected the activities of several TFs such as NF-jB, p53 and Akt. These observations and
456
K. Gohil et al. / Molecular Aspects of Medicine 28 (2007) 453–480
those from other studies show that a single oxidant species can affect the expression of a genome by altering the expression numerous genes that are compartmentalized in distinct regions of the genome. 1.2. Genome-wide responses to AT in vivo Descriptions of vitamin E deficiency symptoms suggest that this micronutrient must have complex effects on the expression of the genome during the mammal’s lifespan, particularly during conception and early development. The two well identified effects of vitamin E deficiency are infertility and neuromuscular disease. More than 80 years after the discovery of the deficiency symptoms, transcriptional signatures of vitamin E for the prevention of infertility are lacking and remain a major deficiency in our knowledge of the molecular actions of AT, the most abundant member of vitamin E family in mammalian tissues. The need for dietary AT for fetal growth may not be unique to rodents as has been suggested by a recent study in humans (Scholl et al., 2006). 1.2.1. Transcriptional responses to dietary AT We are beginning to witness the complexity of the in vivo effects of AT on the neuromuscular and the central nervous systems. Recent genome-wide analyses of mRNAs extracted from muscle and from brain suggest multiple molecular targets that may participate in neuromuscular disease caused by chronic AT deficiency in vivo inflicted by dietary AT-deficiency. In male rats dietary AT-deficiency for 3 months activated 56 muscle genes (Nier et al., 2006); functional classification of the affected genes identified those encoding anti-oxidative, anti-inflammatory, anti-fibrotic and, muscle and extracellular matrix proteins. The transcriptional response to chronic dietary AT deficiency also affects the transcriptomes of the cerebral cortex (Hyland et al., 2006) and the hippocampus (Rota et al., 2005). Dietary AT-deficiency for 14 months in male rats resulted in the repression of 34 genes that included genes encoding myelin proteins and those for neuronal signal propagation, suggesting molecular targets that may account for neurological and electrophysiological alterations (Hyland et al., 2006). AT-sensitive cortical genes also included 11 genes that were induced by AT deficiency and included those encoding catalase and tenascin-R. Nine months of dietary AT-deficiency in male rats affected the transcription of diverse functional clusters of genes that included hormone and hormone metabolism, neuronal survival, dopaminergic system and amyloid proteins in the rat hippocampus (Rota et al., 2005). Dietary AT deficiency also affected the expression of testes genome (Barella et al., 2004b; Rota et al., 2004) and that of the liver (Barella et al., 2004a) and lungs (Oommen et al., 2007). The latter study identified a coordinated regulation of a cluster of 13 cytoskeleton genes that appear to be regulated by serum response factor, a transcription factor important in the development of myocardium and skeletal muscle (Miano, 2003). Genome-wide screens of mRNA expression of T-cells isolated from mice fed different AT-containing diets suggest that AT affects cell cycle and Th1/Th2 balance (Han et al., 2006). These genome-wide
K. Gohil et al. / Molecular Aspects of Medicine 28 (2007) 453–480
457
screens of mRNA expression in vivo suggest that AT has system-wide effects on tissue specific transcriptomes. 1.2.2. Transcriptional responses to AT due to deletion of ATTP-gene The discovery of patients with systemic AT-deficiency in spite of ingesting a diet containing normal amounts of vitamin E has highlighted the physiological requirement of AT for the health of the neurological system (Mariotti et al., 2004; Ouahchi et al., 1995) in humans. The identification of mutations in the ATTP gene which is primarily expressed in the liver has emphasized the importance of a hepatic protein in determining neuromuscular health. Two research groups independently developed transgenic mice lacking ATTP (Jishage et al., 2001; Terasawa et al., 2000). In ATTPKO mice the extrahepatic AT concentrations are
