Endoplasmic reticulum stress induced by aqueous extracts of cigarette smoke in 3T3 cells activates the unfolded-protein-response-dependent PERK pathway of cell survival

Abstract

Cigarette smoke (CS) generally places severe stress on cells, as reflected by gene expression profiling and pathway analysis, which, among other effects, also suggested activation of the unfolded protein response pathway triggered by the stressed endoplasmic reticulum (ER stress). Here, we present data indicating that noncytotoxic concentrations of aqueous extracts of CS induce a distinct ER stress response in immortalized nontransformed Swiss 3T3 cells, primarily by activating the PERK pathway of global protein synthesis inhibition. Activation of PERK and PERK-dependent signaling by aqueous extracts of CS was demonstrated by (i) the inhibition of protein synthesis, (ii) the phosphorylation of PERK and its substrate eIF2α, (iii) the activation of ATF4, and (iv) the expression of ATF4-dependent target genes chop, gadd34, BiP, and atf3. Within the dose range tested, all effects appeared to be transient in nature, while the periods of recovery from ER stress were clearly concentration dependent. In contrast to these data and to the effects seen with thapsigargin (used as positive control), only minor effects were observed for the activation of xbp-1, a common target of the other two canonical sensors of ER stress, i.e., ATF6 and IRE1. In mechanistic terms, neither the disruption of energy levels nor a contribution of arylating quinones played a major role under the experimental conditions tested. Notably however, the effects of aqueous extracts of CS on the ER could be mimicked in the presence of acrolein at CS-relevant concentrations, indicating that CS interferes with proper ER function, presumably due mainly to changes in cellular redox homeostasis. Since ER stress has been linked to diseases that are also related to CS exposure, these data are relevant in the discussion of a general molecular mechanism of CS-induced disease.

Introduction

Although cigarette smoke (CS) is a causal component of a multitude of severe, mostly inflammation-related diseases such as cancer, cardiovascular diseases, and chronic obstructive pulmonary disease (for reviews see [1], [2]), the underlying fundamental mechanisms are only just beginning to emerge. The main difficulty in elucidating the molecular and cellular modes of action of CS-induced adverse health effects is clearly related to the complex nature of CS, which is thought to be composed of up to 5000 chemicals. Thus, it is not surprising that current data delineate a complex picture of cellular and tissue responses to CS exposure, which, in general, are characterized by the cell's attempt to cope with and adapt to the impaired intracellular and microenvironmental conditions (summarized in a series of reviews; see [3]); [4].

In an attempt to further characterize the fundamental mechanisms of CS-dependent molecular and cellular toxicity, our laboratory recently implemented gene expression profiling (microarray) studies, which revealed a distinct transcriptional response in CS-exposed cellular [5] and rodent respiratory tissue material [6], [7]. Collectively the resulting expression profiles are governed by the differential regulation of xenobiotic and metabolism (phase I)-, antioxidant and phase II-, and inflammation-related genes. However, they also provide evidence for the activation of other mechanistic responses, an intriguing example of which is represented, especially in vitro [5], by the upregulation of genes indicative of a compromised endoplasmic reticulum (ER) function, such as gadd34, chop, and atf3 (marked in red in Fig. 1). Because this response could further explain previous isolated data on a reduced translational efficiency in cells exposed to aqueous extracts of CS [8], we sought to follow up on this by mechanistically investigating cells for potential ER stress by exposure to aqueous extracts of CS.

Attenuation of protein biosynthesis activities is an adaptive response evolved by eukaryotic cells in response to various types of stress affecting cellular homeostasis to cope with conditions such as facing the accumulation of misfolded and unfolded proteins, amino acid deprivation, and perturbations of the redox and/or energy status not permitting proper ER function, all of which are generally subsumed under the term “ER stress” (recently reviewed in [9]). This mechanism, the unfolded protein response (UPR), couples the ER protein folding load with the ER protein folding capacity and is generally regarded a cellular tool to successfully combat ER stress [10], [11]. Basically, UPR signaling in mammalian cells is orchestrated by three different arms (Fig. 1), each of which is initialized by a distinct sensor anchored in the ER as transmembrane protein and termed as PERK, IRE1, and ATF6 [9]. All three transducers of ER stress are activated by a similar mechanism involving the ER stress stimulus, i.e., an overload of unfolded "client" proteins, which induces the release of the ER chaperone BiP from complexes with PERK, IRE1, and ATF6. Most probably, this signal finally leads to spontaneous homodimerization and trans-autophosphorylation of PERK and IRE1, whereas ATF6 is cleaved to activation in the Golgi [9]. UPR signaling through activated IRE1 and ATF6 eventually results in the expression of a specific subset of genes regulated by ER-stress-sensitive consensus sites (i.e., ER stress response element (ESRE) and unfolded protein response element (UPRE)) present in their promoter control regions. An essential component of UPR-dependent IRE1 and ATF6 signaling is xbp-1 (X-box DNA-binding protein 1), which is transcriptionally induced by ATF6. The resulting transcript (xbp-1^u) has to be processed by the RNase activity inherent to the cytosolic domain of IRE1, removing a 26-nucleotide intron to yield mature xbp-1^s mRNA (for review, see [12]). Whereas the unspliced variant of XBP1 had been reported to be labile and even to repress UPR target genes [13], the spliced version of XBP1 has been an efficient transcriptional activator of ESRE/UPRE-responsive genes [9]. The most immediate response to ER stress, however, is provoked by the third signaling component of the UPR, i.e., the Ser/Thr kinase PERK [11], [14]. Upon activation, PERK phosphorylates the eukaryotic translation initiation factor 2α (eIF2α) on Ser 51, resulting in an inhibition of 80S ribosome assembly and, consequently, an abrupt downregulation of general protein synthesis. However, phosphorylated eIF2α selectively stimulates the translation of a specific subset of genes characterized by the presence of multiple open reading frames (ORFs) in the 5′ upstream region of the corresponding mRNAs, a structural feature that permits proper translation only in the presence of impaired ribosome formation. A major target of this striking bypass mechanism is represented by the transcription factor ATF4 [15], which is an efficient inducer of UPR-responsive genes [16]. In addition to ATF4 expression, activated PERK has also been reported to directly activate the transcription factor Nrf2 [17], which is a strong inducer of antioxidant and phase II-related genes and a major target of CS exposure in vitro [18] and in vivo [19], thus enhancing its prosurvival effects (Fig. 1).

Here, we report that aqueous extracts of CS activate primarily the UPR-dependent PERK signaling pathway of protein synthesis inhibition, resulting in ATF4 activation and ATF4-specific gene expression in nontransformed Swiss 3T3 fibroblasts, a cell line used extensively in our laboratory to characterize the molecular and cellular effects induced by aqueous extracts of CS ([3] and references cited therein). Our presented results indicate that the CS gas phase component acrolein (2-propenal) is a major culprit in ER stress induced by aqueous extracts of CS.