Smoke-induced microRNA and related proteome alterations. Modulation by chemopreventive agents



Dysregulation of microRNAs (miRNAs) has important consequences on gene and protein expression since a single miRNA targets a number of genes simultaneously. This article provides a review of published data and ongoing studies regarding the effects of cigarette smoke (CS), either mainstream (MCS) or environmental (ECS), on the expression of miRNAs and related proteins. The results generated in mice, rats, and humans provided evidence that exposure to CS results in an intense dysregulation of miRNA expression in the respiratory tract, which is mainly oriented in the sense of downregulation. In parallel, there was an upregulation of proteins targeted by the downregulated miRNAs. These trends reflect an attempt to defend the respiratory tract by means of antioxidant mechanisms, detoxification of carcinogens, DNA repair, anti-inflammatory pathways, apoptosis, etc. However, a long-lasting exposure to CS causes irreversible miRNA alterations that activate carcinogenic mechanisms, such as modulation of oncogenes and oncosuppressor genes, cell proliferation, recruitment of undifferentiated stem cells, inflammation, inhibition of intercellular communications, angiogenesis, invasion, and metastasis. The miRNA alterations induced by CS in the lung of mice and rats are similar to those observed in the human respiratory tract. Since a number of miRNAs that are modulated by CS and/or chemopreventive agents are subjected to single nucleotide polymorphisms in humans, they can be evaluated according to toxicogenomic/pharmacogenomics approaches. A variety of cancer chemopreventive agents tested in our laboratory modulated both baseline and CS-related miRNA and proteome alterations, thus contributing to evaluate both safety and efficacy of dietary and pharmacological agents.

The discovery of microRNAs (miRNAs) has opened new avenues in a variety of disciplines related to biomedical research. miRNAs constitute a family of small noncoding and single-stranded RNAs, formed of 18 to 25 nucleotides, which modulate gene expression either by translational inhibition or by exonucleolytic messenger RNA (mRNA) decay.1,2 Thus, miRNAs provide a major epigenetic mechanism that regulates translation of expressed genes into proteins. Analysis of miRNAs compares favorably with analysis of other postgenomic end-points in terms of relationship with the phenotype. In fact, the information provided by transcriptome analyses is redundant due to the fact that mRNA can be regulated post-transcriptionally. Indeed, only less than 6% of mRNA reaches the ribosomes.3 On the other hand, a very tiny proportion of proteins can be analyzed by means of the available techniques, which are just able to analyze hundreds of the 1.5 million proteins composing the cell proteome.4 In contrast, microarray analyses provide a broad coverage of miRNA domain, since miRNAs regulate a number of genes simultaneously by complementary binding of the 3'-untranslated region of target mRNAs, a single miRNA targeting approximately 200 transcripts simultaneously.5 The miRBase Version 16.0 has 1,048 miRNA sequences annotated in the human genome. It has been estimated that miRNAs may regulate up to 30% of all protein-coding genes in the human genome.6 Accordingly, dysregulation of miRNAs can have an important consequence in dysregulation of genes, and miRNA analysis can provide a more comprehensive understanding of lung cancer pathogenesis.7

miRNAs interact with classic oncogene and tumor suppressor networks thereby contributing to the initiation and progression of many, if not all, human malignancies.8 In general, miRNA expression is downregulated in cancer as compared with normal tissues.9 Downregulation of specific miRNAs in lung tumors has been observed in both mice and humans.10,11 Less attention has been paid to miRNA and proteome changes occurring in apparently healthy tissues of subjects exposed to cigarette smoke (CS), either mainstream (MCS), which is inhaled by active smokers, or environmental (ECS), which is inhaled by passive smokers.

The first goal of the present article was to provide an overview of studies evaluating dysregulation of miRNA expression and proteome profiles in mice, rats, and humans exposed to CS. A second goal was to review studies investigating modulation by chemopreventive agents of CS-related alterations induced by either ECS or MCS in animal model. Some of the data generated in our laboratory are already published, whereas other data were so far unpublished.

Interplay Between miRNAs and Other Smoke-Related Intermediate Biomarkers in Rats

MiRNA alterations are not the only molecular event in response to CS. Therefore, their biological meaning has to be interpreted in connection with other end-points. As an example, Figure 1 shows a series of molecular alterations detected in the respiratory tract of Sprague-Dawley rats after 4 to 5 weeks of whole-body exposure to ECS. They include: (i) bulky adducts to nuclear DNA, detected by 32P postlabeling in bronchoalveolar lavage (BAL) cells, dissected tracheal epithelium, and lung homogenates. The panel shows the time-course formation of DNA adducts during the first 5 weeks of exposure and its reversibility upon discontinuation of exposure to ECS during the last week of exposure.12,13 In contrast, after 9 months of exposure, bulky DNA adducts in the lung of either wild-type or P53 mutant mice did not decrease upon discontinuation of exposure to ECS for 1 week;14 (ii) induction of oxidative DNA damage, detected by 32P postlabeling measuring 8-hydroxy-2'-deoxyguanosine (8-oxo-dGuo) in lung homogenates;13,15 (iii) stimulation of apoptosis in pulmonary alveolar macrophages (PAM) and bronchial epithelium, as evaluated by TdT-mediated dUTp nick end labeling (TUNEL) method;13,16 (iv) stimulation of proliferation in PAM, as assessed by measuring proliferating cell nuclear antigen (PCNA) by immunohistochemistry;13,16 (v) loss of Fhit gene and protein in PAM and bronchial epithelium, as evaluated by RT-PCR, QPCR, immunohistochemistry, and Western blot analyses;17 (vi) induction of cytogenetical damage in PAM, as evaluated by measuring the frequency of polynucleated (PN) and micronucleated (MN) PAM;15 (vii) dysregulation of miRNA expression in lung homogenates, as evaluated by miRNA microarrays, with 126 of 484 analyzed miRNAs (26.0%) downregulated by ECS;18,19 (viii) changes of multigene expression in lung homogenates, as evaluated by cDNA microarrays, with 107 of 4,854 miRNas (2.9%) upregulated by ECS;20 (ix) changes of proteome profiles in lung homogenates, as evaluated by antibody microarrays, with 50 of 518 proteins (9.7%) upregulated by ECS.21,22

Figure 1.

Molecular alterations in the respiratory tract of Sprague-Dawley rats exposed to ECS. ECS-induced miRNA downregulation (g) is the link between DNA damage (a, b, f), transcriptional changes (h), and postgenomic alterations, such as Fhit loss (e), induction of apoptosis (c), cell cycle stimulation (d), and upregulation of protein expression (i). See text for details. [Color figure can be viewed in the online issue, which is available at]

Thus, within the respiratory tract of rats, ECS caused an early DNA damage in cells (PAM), tissues (tracheal epithelium), and organs (lung), as shown by an increase of bulky DNA adducts, oxidative DNA damage, cytogenetical alterations and loss of the oncosuppressor gene Fhit. DNA damage was accompanied by a massive downregulation of miRNas in lung, which as expected resulted in an extensive upregulation of both mRNAs and proteins. Phenotypically, a stimulation of cell apoptosis and proliferation was observed (see Figs. 1c and 1d). These findings were further confirmed by multigene expression analyses in A/J mice, either wild-type or P53 mutant,23 SKH-1 mice,24 and Sprague-Dawley rats,21 which showed dramatic alterations of genes involved in a variety of functions, including apoptosis, cell proliferation, protein repair, stress response, angiogenesis, inflammation, and xenobiotic metabolism. These alterations bear relevance not only in carcinogenesis but also in other CS-related diseases. For instance, alterations of PKC and other inflammation-related activities play a role in asthma and chronic bronchitis,23 trypsin-antitrypsin balance plays a role in emphysema,23 and alterations of plasminogen activator inhibitors play a role in thrombosis.24 It is also of interest that CS upregulates MDR1 expression, which suggests that MDR-related resistance to anticancer drugs occurs since early stages of lung carcinogenesis.24 Similar gene expression alterations were found in the bronchial epithelium of smoking humans.25

Exposure of Sprague-Dawley rats to ECS for 4 weeks dramatically altered the pulmonary miRNA expression, as evaluated by microarray and qPCR.18 In particular, the most remarkably downregulated miRNAs were those of the let-7 family, miR-10, miR-26, miR-30, miR-34, miR-99, miR-122, miR-123, miR-124, miR-125, miR-140, miR-145, miR-146, miR-191, miR-192, miR-219, miR-222, and miR-223. These miRNAs regulate stress response, apoptosis, proliferation, angiogenesis, and expression of genes. In contrast, miR-294, an inhibitor of transcriptional repressor genes, was upregulated by ECS.18

When evaluating the data generated in experimental animals, a major problem is their translability to humans. Since miRNAs are highly conserved during phylogenesis, in principle it would be expected that the interspecies translability of miRNA data is appropriate. Dealing with CS-induced miRNA alterations, there was a strong parallelism between rodent miRNAs and their human homologues, which are often transcribed from genes localized in fragile sites deleted in lung cancer.18 In particular, miR-125a-prec and miR-125b were two of the most strikingly ECS-downregulated human miRNA homologues. Interestingly, miR-125b is recognized as the homologue of Caenorhabditis elegans lin-4, the first miRNA discovered.26 MiR-125 plays an important role in lung carcinogenesis, its genetic targets including the ERBB2 proto-oncogene that encodes for the EGF receptor, which is highly expressed in carcinomas.27 miR-125 genes are located in 11q23-q24 and 21q11.1, regions that are frequently deleted in lung cancer.28

A critical analysis of translability to humans of the results obtained in CS-exposed rodents has demonstrated that alterations of miRNA expression profiles, as detected by independent groups, are comparable in mice, rats and humans.29

Interestingly, the human homologues of five miRNAs downregulated by ECS, including let-7a, miR-124, miR-125, miR-140, and miR-146, are characterized by functionally relevant genetic polymorphisms, which may be involved in explaining the interindividual variability in susceptibility toward ECS.18 In fact, SNPs in the let-7a complementary site in k-Ras have been shown to increase the risk of nonsmall cell lung cancer.30 However, genetic variants of miRNA sequences barely affect survival of patients bearing this type of cancer.31

It is important to explore their consequences on miRNA-targeted proteins at the phenotypic level in order to shed light on the biological meaning of CS-related miRNA alterations. To this respect, it is noteworthy that the whole-body exposure of Sprague-Dawley rats to ECS did not significantly decrease the levels of any protein tested in lung S12 fractions, as evaluated by antibody microarrays and confirmation of selected proteins by Western blot. In contrast, 50 proteins were increased more than twofold.22 As shown in Figure 2, the upregulated proteins belonged to a variety of functional categories, all of which are known to be involved in pulmonary carcinogenesis [see Ref. 19 and the references cited therein]. In particular, among the stress response proteins, the phosphoprotein Fos is the product of the proto-oncogene c-Fos that forms the activator-protein-1 (AP-1) transcription factor by binding the c-jun product. The ezrin-radixin-moesin-binding phosphoprotein 50 (EBP50) regulates the secretory activity of the nonciliated, nonmucous Clara cells. The α, β, and γ subunits of the NF-κB inhibitor allow NF-κB to translocate into the nucleus, whereas heme oxygenase is an important defense mechanism against smoke-related oxidative damage. ECS had limited effects on proteins involved in DNA repair, which suggests a poor inducibility of DNA repair genes by CS. Four of the eight proteins classified in the category “protein repair/removal” are components of the ubiquitin/proteasome system, a major pathway for protein degradation. The ECS-stimulated proteins involved in cell replication included several cyclins and cyclin-dependent kinases. The observed stimulation of PCNA is consistent with the increase of this protein in the bronchial epithelium. In lung S12 fractions, ECS did not significantly enhance P53 protein, which tends to accumulate in the nucleus, but increased both P21, the main P53 effector for cell cycle arrest, and 53BP2. In addition, ECS increased the amounts of seven proteins playing a proapoptotic role, among which four caspases. Finally, ECS stimulated proteins related to phagocytosis, endocytosis, and intracellular vesicular traffic, and proteins related to the immune response, such as macrophage-associated receptors, leukocyte-associated antigens, and the T-cell protein FYB. These findings correlate well with the conclusions of previous studies evaluating multigene expression in rodents.22

Figure 2.

Functional categories of proteins upregulated in the lung of rats exposed to ECS in parallel to miRNA downregulation. The data are from Izzotti et al.22

Further studies on smoke-related proteomic alterations were performed by Zhang et al.32 by testing the lung tissue from Wistar rats exposed to CS. About 500 protein spots were detectable by two-dimensional electrophoresis. Among the 28 differentially expressed protein spots, 18 were identified by mass spectrometry. Some of them (enolase, dimethylargimine, dimethylaminohydrolase 1, iodothyronine-5'-monodeiodinase, C44B7.7, and one unnamed protein) were upregulated by CS.

Alteration of miRNA Expression Profiles in Mice Exposed Either to ECS or MCS

We implemented a series of studies aimed at evaluating whether age affects miRNA alterations induced by CS in mouse lung. In the earlier study of this series,33 we first evaluated the baseline miRNA expression in the mixed-cell population of the lung from Swiss ICR(CD-1) mice, as related to different developmental stages during postnatal life, from birth to adulthood. Eleven miRNAs, 5 of which are involved in embryogenesis and morphogenesis, varied significantly from birth to the end of the weanling period. These variations are linked to perinatal stress in the lung and to maturation of this organ in terms of tissue growth and remodeling, cell differentiation, and physiological vascularization. The changes in baseline miRNA expression tended to become less attenuated during the shift to adulthood. This physiological situation was altered by CS. In fact, exposure of either postweanling or adult mice to ECS resulted in dramatic changes in the miRNA domain, mainly in the sense of downregulation. Some of the ECS-dysregulated miRNAs are involved in stress response (miR-122a, miR-124a, and miR-125b), with special reference to NF-κB (miR-30b), protein repair (miR-30b and miR-431), and apoptosis (miR-99b), consistently with the evidence that ECS is a potent inducer of apoptosis in the respiratory tract.

Thus, part of the observed changes in miRNA expression can be interpreted as adaptive mechanisms aimed at attenuating the effects of ECS, whereas other changes, such as let-7 downregulation, are likely to be involved in the pathogenesis of ECS-related lung cancer via both epigenetic and genotoxic mechanisms. In general, miRNA alterations were more evident in postweanling mice exposed to ECS since birth than in their dams exposed to ECS during adulthood. These studies also showed that alterations in miRNAs involved in stem cell maintenance and recruitment, such as miR-124, miR-214, miR-335, miR-376, and miR-411, are induced by ECS only in young mice.33 These miRNA alterations are likely to be related to changes in the composition of the cell population occurring early in life. Indeed, exposure to ECS resulted in pulmonary stem cell recruitment only in postweanling mice and not in adult mice.34 Accordingly, in agreement with human data,35 these findings suggest that CS reverses lung tissue differentiation. It should be noted that stem cells have been demonstrated to be more susceptible to DNA damage induced by polycyclic aromatic hydrocarbons than their differentiated counterparts.36

Further on, we evaluated the effects of varying doses of MCS on miRNA alterations in the lung of mice exposed since birth for 4 weeks.37 The MCS-induced downregulation of miRNAs was dose-dependent. However, miRNAs appeared to be a less sensitive target than bulky DNA adducts and 8-oxo-dGuo in revealing exposure to MCS at low doses, which suggests the existence of thresholds in the induction of adverse effects on the miRNA machinery. An exception was miR-30, involved in stress response and NF-κB activation, which was downregulated even at the lowest MCS dose tested. The time-course analysis of the expression of 697 miRNAs showed that MCS-induced miRNA dysregulation was considerably attenuated 1 week after smoking cessation. These general miRNA kinetics are well exemplified by the let-7 miRNA family, the expression of seven let-7 miRNAs being completely restored 1 week after smoking cessation. However, the recovery of the expression of other miRNAs was incomplete after 1 week, which suggests that a longer time period is needed to fully restore the expression of these miRNAs. For instance, these patterns were observed with miR-34b, which plays an important role as a P53 effector, due to the fact that P53 directly induces mir-34 expression,38 and upregulation of this miRNA restores P53 functions in cancer cells.39 Our data suggest that the P53 pathway requires more than 1 week to be fully restored after smoking cessation. Other miRNAs showing the same trend were miR-345, miR-421, and miR-450, which recognize as molecular targets genes involved in the Ras oncogene pathway, playing a fundamental role in lung carcinogenesis.37

We evaluated miRNA expression in the lung of mice affected by various histopathological alterations induced by long-term exposure to MCS.40 As reported in Figure 3, miRNAs involved in inflammation and angiogenesis were dysregulated in all histopathological situations examined. In pneumonia, miR-29c upregulation was related to P53 overexpression, while in both adenoma and carcinoma downregulation of a variety of miRNAs resulted in P53 silencing.40 Both adenoma and carcinoma were characterized by dysregulation of miRNAs involved in the loss of intercellular adhesion, cell proliferation, and MDR1 activation. However, the number of altered miRNAs involved in cell proliferation was by far more relevant in carcinoma than in adenoma, and only carcinoma was characterized by downregulation of k-Ras silencing miRNAs. This finding might potentially explain why, despite the early occurrence of k-Ras mutations in mice exposed to CS,41 cancer occurs only at late stages of the carcinogenesis process, when the miRNA machinery is irreversibly altered. In advanced carcinoma, this situation is accompanied by the silencing of miR-162 that inhibits Foxo1, a tumor suppressor gene inhibiting cancer cell proliferation and spread of metastasis.40

Figure 3.

MCS-induced miRNA alterations in mice, as related to lung histopathology.

Dysregulation of miRNAs in the Airway Epithelium of Smoking Humans

Schembri et al.11 compared miRNA expression in the bronchial airway epithelium from 10 never smokers and 10 current smokers by using a battery of 467 miRNAs analyzed by microarrays. A total of 28 miRNAs were found to be differentially expressed in smokers, 23 of which being downregulated and 5 being upregulated as compared with never smokers. The most extensively altered miRNA was miR-218, transcribed within the intronic region of the oncosuppressor gene SLIT2, which is frequently inactivated in lung tumors. MiR-218 was downregulated not only in smokers but also in human lung carcinoma and bronchial epithelial cells treated in vitro with a CS condensate. MiR-218 transfection of epithelial bronchial cells corrected CS-induced miRNA downregulation and inhibited the gene expression response activated by the transcription factor MAFG.11 Three additional miRNAs (miR-15a, miR-199b, and miR-125b) that were downregulated in smokers have also been shown to be downregulated in lung cancer, thus suggesting that downregulation of miRNAs in the bronchial epithelium of smokers could be related to the development of tobacco-related cancers.11,29

Downregulation of miRNAs was also observed by analyzing 60 biopsies of smokers with normal or hyperplastic appearance as compared with never smokers. A total of 69 miRNAs were differentially expressed in smokers and nonsmokers. Moreover, miR-32 and miR-34c progressively downregulated their expression during the change from normal bronchial tissue to the appearance of squamous cell carcinoma.35

Chemoprevention of ECS-Related miRNA and Related Proteome Alterations in Rats

Table 1 shows a list of the cancer chemopreventive agents that have been investigated or are being investigated in our laboratory for the ability to modulate CS-induced miRNA alterations in rodent tissues. They include agents, either natural or synthetic, that were administered with the diet to both CS-free and CS-exposed rodents. This dual approach has the goal of evaluating not only the efficacy of chemopreventive agents in counteracting CS-induced miRNA alterations but also their safety, as inferred from the effects on baseline miRNA expression profiles.

Table 1. Chemopreventive agents evaluated in our laboratory for the ability to modulate smoke-induced alterations of miRNAs in rodents
inline image

As indicated in Table 1, five agents, including N-acetylcysteine (NAC), oltipraz (OPZ), β-naphthoflavone or 5,6-benzoflavone (5,6-BF), phenethyl isothicyanate (PEITC), and indole-3-carbinol (I3C) and 2 combinations thereof (NAC + OPZ and PEITC + I3C) were administered to adult Sprague-Dawley rats either with the diet or with the drinking water. These agents had previously been evaluated for the ability to inhibit the ECS-induced formation of bulky DNA adducts in BAL cells, tracheal epithelium, lung, and heart, oxidative damage (8-oxo-dGuo) to lung DNA, hemoglobin adducts, cytogenetical damage in PAM and bone marrow erythrocytes,15 and apoptosis in PAM and bronchial epithelium.16 Further on, we evaluated the ability of the same agents to inhibit the ECS-related formation of DNA adducts in both lung and liver and to modulate the expression of 4,858 genes in the lung of mice, either unexposed or exposed to ECS. Excepting OPZ, all treatments inhibited the formation of bulky DNA adducts in the lung of ECS-exposed rats by at least 50%. Hierarchical cluster analysis and principal component analysis (PCA) allowed us to classify the agents according to their influence on baseline gene expression and their ability to attenuate ECS-induced transcriptome alterations. PEITC and I3C were the most effective but at the same time were the least safe agents; 5,6-BF displayed intermediate patterns. NAC had a medium efficacy and was the safest agent. OPZ was poorly effective in lung and considerably altered the baseline gene expression in liver.21

Microarray data relative to 484 miRNAs and quantitative PCR data in the lung of the same rats were subjected to scatterplot, hierarchical cluster analysis, and PCA. The results of hierarchical cluster analyses showed that none of the above chemopreventive regimens appreciably affected the baseline miRNA expression, indicating potential safety. On the other hand, all of them attenuated ECS-induced alterations to a variable extent, indicating potential preventive efficacy. Likewise, as shown in Figure 4, the results of PCA showed that the overall miRNA expression profiles in ECS-exposed rats fell far away from those of sham-exposed rats, in two opposite quadrants. All chemopreventive agents, in the absence of exposure to ECS, fell in the same quadrant where Sham was allocated, thus reflecting the poor alterations of miRNA expression exerted by all chemopreventive regimens tested. On the other hand, when given to ECS-exposed rats, all chemopreventive agents tended to depart from ECS and to approach Sham. In the case of the combined treatment with PEITC and I3C, however, PCA allocation was the farthest away from ECS, but at the same time it was far away from Sham, thus reflecting a nonphysiologic modulation of ECS-related miRNA expression. The main ECS-altered functions that were modulated by chemopreventive agents included mechanisms involved in cell proliferation, apoptosis, differentiation, k-Ras activation, P53 functions, NF-κB pathway, transforming growth factor-related stress response, and angiogenesis. We refer to the original article for more details.19

Figure 4.

Bidimensional PCA showing the overall expression profiles of 484 miRNAs in rat lung, as related to exposure to ECS and treatment with chemopreventive agents. Modified with permission from Izzotti et al., Cancer Prev Res, 2010, 3, 62-72, © American Association for Cancer Research.

In order to evaluate the consequences of miRNA alterations at the phenotypic level, the lungs from four groups of adult Sprague-Dawley rats that had been used for evaluating a variety of biomarkers, among which multigene expression and miRNA expression (see above), were subjected to proteome analysis.22 They included (i) rats kept in filtered air (sham), (ii) rats treated with NAC with the drinking water, (iii) rats exposed to ECS for 4 weeks, and (iv) ECS-exposed rats treated with NAC. Evaluation of 518 proteins in rat lung by antibody microarray and confirmation of selected proteins by Western blot showed that 50 proteins (9.7%) were increased more than twofold following exposure to ECS. None of the 518 tested proteins was significantly altered by NAC administration to ECS-free rats, while 42 proteins (8.1%) were still increased more than twofold in ECS-exposed rats treated with NAC. The proteins that were normalized by NAC in the lung of ECS-exposed rats included proteins involved in stress response and NF-κB activation (Fos, Iκkα, Iκkβ, and Iκkγ), positive regulators of the cell cycle (cyclins A and F), and a proapoptotic protein (caspase-14).22 These data are consistent with the mechanisms of this thiol, a GSH precursor and analogue, whose activity as a nucleophile and scavenger of reactive oxygen species (ROS) includes the ability to inhibit the apoptotic process resulting from generation of ROS and imbalances in the redox potential.43,44 NAC has been shown to prevent the tumors and other histopathological alterations induced by MCS early in life when given under experimental conditions simulating exposure to a current smoker45 or even when given prenatally.42,46

Interestingly, by comparing gene expression and proteome data for 278 activities that were evaluated for both end-points, there were significant correlations between these two end-points in all four experimental groups, including sham-exposed rats (r = 0.134, p < 0.05), NAC-treated rats (r = 0.123, p < 0.05), ECS-exposed rats (r = 0.133, p < 0.05), and ECS-exposed rats treated with NAC (r = 0.117, p = 0.05).22

Chemoprevention of miRNA Alterations in Neonatal Mice Exposed Either to ECS or MCS

A series of chemopreventive agents (see Table 1) were tested or are being tested in our laboratory for the ability to modulate miRNA expression in mouse tissues and, in parallel, for prevention of CS-induced tumors. The general design of these studies involves exposure of mice to CS for 4 months, starting soon after birth. The chemopreventive agents are administered with the diet after weanling (about 4 weeks). After 2 to 6 additional weeks, groups of mice are sacrificed for the analysis of miRNAs and other intermediate biomarkers. The remaining mice are kept alive until 7 to 11 months of age, when they are sacrificed in order to evaluate tumors and other histopathological alterations. The rationale for starting exposure at birth is that, due to the hyperoxic challenge in the neonatal lung during the sudden transition from the maternal-mediated respiration to the autonomous pulmonary respiration, the mouse lung undergoes at birth “physiological” genomic alterations, such as DNA adducts and oxidative DNA damage.20 Although these alterations tend to be compensated by overexpression of genes attenuating stress and oxidative damage,20 the perinatal period is a critical moment of life. In addition, several other mechanisms contribute to the high susceptibility to carcinogens when exposures occur early in life (see Ref. 47 and the references cited therein). Neonatal mice are highly susceptible to a variety of molecular, biochemical, and cytogenetical alterations induced by ECS.48 Starting exposure to MCS at birth, a remarkable carcinogenic response is detectable after 7 to 8 months,47 which is much more evident than after exposure to adult mice under the same experimental conditions.49 Although less potently, a significant carcinogenic response and a variety of histopathological alterations are also detectable in the lung and other organs of ECS-exposed mice.50

Based on these premises, we evaluated modulation of miRNA expression by budesonide and PEITC in both lung and liver of Swiss ICR (CD-1) mice, either unexposed or exposed to ECS since birth. Thus, information was provided regarding miRNA expression as related to (i) physiological interorgan differences, (ii) dysregulation following exposure to ECS since birth, (iii) effects of the tested chemopreventive agents on the baseline miRNA expression, and (iv) modulation by these agents of ECS-related alterations, under a situation mimicking an intervention in passive smokers. In the comparison between lung and liver from smoke-free mice it should be taken into account that both organs are composed of multiple cell types. The general trend was toward a higher baseline expression of miRNAs in liver than in lung, presumably due to the multiplicity and complexity of liver functions. The most abundant miRNA in the liver, miR-122, is known to be involved in cholesterol biosynthesis and maintenance of the liver phenotype. As in the previously discussed study,33 dysregulation of miRNA expression in the lung of ECS-exposed mice was mainly oriented in the sense of downregulation of a variety of miRNas involved in important cellular functions. Even more intense was the ECS-related dysregulation, both in the sense of upregulation and downregulation, of those liver miRNAs that are mainly involved in adaptive functions.

In the lung of smoke-free mice, budesonide had negligible effects on miRNA expression profiles, and PEITC downregulated five miRNAs only. In contrast, both agents induced profound alterations of miRNA expression in liver, mostly in the sense of downregulation, which reflects adverse effects in the liver, contrasting with the potential safety and efficacy of these agents in the lung. This interorgan difference was maintained in ECS-exposed mice. In fact, budesonide and, more efficiently, PEITC tended to attenuate ECS-related miRNA alterations in lung, although both agents failed to fully restore the physiological situation. In the liver, PEITC and especially budesonide exhibited a poor ability in counteracting the dysregulation of miRNA expression induced by ECS.51 Attenuation by PEITC and budesonide of early ECS-related alterations of miRNA expression in the lung correlate with the protective effects observed in the same groups of mice, at 11 months of life, towards ECS-induced histopathological alterations, including lung tumors.52 Moreover, both PEITC and budesonide were quite effective in inhibiting the formation of lung tumors in Swiss H mice exposed to MCS, at 7 months of life, when the agents were administered according to a protocol mimicking an intervention in current smokers. Interestingly, PEITC lost part of its cancer chemopreventive activity when administered according to a protocol mimicking an intervention in ex-smokers, while the protective capacity of budesonide was unchanged. These differential patterns are consistent with the mechanisms of action of these agents. In fact, PEITC mainly works at the metabolic level, while the glucocorticoid budesonide is a potent antiinflammatory agent and is therefore expected to exert protective effects also in advanced carcinogenesis stages.45

A further extensive study involved both miRNA and related proteome analyses in 20 experimental groups, for a total of 415 ICR (CD-1) mice. The investigated agents included myo-inositol, suberoylanilide hydroxamic acid (SAHA or Vorinostat), bexarotene (or Targretin), and pioglitazone, each of them tested at four dose levels in smoke-free mice, starting after weanling (4 weeks) and continuing for an additional 6 weeks. Moreover, bexarotene and pioglitazone were also tested at the maximum nontoxic dose in mice exposed to MCS since birth until the end of the experiment. The results relative to modulation by these agents of miRNA expression in the lung of mice, either smoke-free or MCS-exposed, are shown in Table 2.

Table 2. Modulation by myo-inositol, vorinostat, bexarotene, and pioglitazone of miRNA expression in the lung of ICR(CD-1) mice, either smoke-free or exposed to MCS
inline image

We analyzed proteome profiles by antibody microarray evaluating the expression of 1,164 proteins, with particular reference to those targeted by miRNAs modulated by chemopreventive agents. Myo-inositol, administered at the doses of 8, 4, 2, and 1 g/kg diet, affected the baseline expression of 9 out of 694 mouse-specific miRNAs, 6 of which were upregulated and 3 were downregulated. In fact, upregulation of miR-302 correlated with overexpression of cyclin-dependent kinase inhibitor 1A (p21), which is involved in the cell cycle regulation; upregulation of miR-509 correlated with overexpression of Rho GTPase activating protein, which is involved in metabolic functions; upregulation of miR-880 correlated with overexpression of serine/threonine kinase 24, which is involved in stress response; and downregulation of miR-543 correlated with a reduced expression of both epidermal growth factor receptor pathway substrate 8, which is involved in cell differentiation, and of nuclear receptor subfamily 3, which is involved in inflammatory responses, cellular proliferation and differentiation.

Vorinostat, administered at the doses of 1,000, 500, 250, and 125 mg/kg diet, altered the baseline expression of seven miRNAs, four of which were upregulated and three were downregulated. Again, miRNA data were consistent with proteome data. First of all, consistently with the main mechanism of this histone deacetylase inhibitor, upregulation of miR-125 and miR-466 correlated with overexpression of histone deacetylase 3, which regulates epigenetic expression; upregulation of miR-302 correlated with overexpression of cyclin-dependent kinase inhibitor 1A (P21), a P53-dependent negative regulator of the cell cycle; upregulation of miR-183 correlated with overexpression of both integrin beta-1, which regulates cell adhesion, and of protein phosphatase 2, which is implicated in the negative control of cell growth and division; and upregulation of miR-684 correlated with a reduced expression of importin beta-1, which plays a role in signal transduction.

Both bexarotene and pioglitazone, administered to smoke-free mice at the doses of 240, 120, 60, and 30 mg/kg diet, did not affect the baseline miRNA expression profiles. At the maximum doses tested, both nuclear receptor agonists counteracted the MCS-related downregulation of a few miRNAs. In particular, bexarotene prevented downregulation of miR-493, which correlates with upregulation of two proteins targeted by this miRNA, including keratin pan, which is involved in cell differentiation, and cell division cyclin 27, which is involved in the cell cycle regulation. Pioglitazone counteracted the MCS-related downregulation of miR-218, miR-296, miR-305, and miR-764. In parallel, proteome analysis showed that upregulation of miR-218 correlated with overexpression of thioredoxin and Ras proteins; upregulation of miR-335 correlated with overexpression of insulin-like growth factor; and upregulation of miR-764 correlated with overexpression of peroxisomal D3,D2,enoyl-CoA isomerase.


The results of the herein reported studies concur to the conclusion that exposure to CS, either MCS or ECS, causes an intense dysregulation of miRNA expression profiles in the respiratory tract of mice, rats, and humans. miRNA alterations were dose-dependent and were more frequently oriented in the sense of downregulation. They were accompanied by upregulation of both gene expression and protein expression. In some studies it was possible to evaluate the parallelism between CS-related miRNA changes and expression of the targeted proteins. In any case, due to the previously discussed ability of each miRNA to target a number of genes simultaneously, evaluation of miRNA expression profiles appears to provide a more sensitive tool than evaluation of either multigene or protein expression. For instance, the 26.0% of 484 analyzed miRNAs were downregulated in the lung of rats exposed to ECS, whereas the 2.9% of 4,858 genes and the 9.7% of 518 proteins were upregulated.18,22

Dysregulation of miRNA expression following exposure to CS has to take into account that CS is a complex mixture containing thousands of chemicals that act through a variety of different mechanisms, which renders the interpretation of these data more complicate and variegate. On one hand, the trend to a CS-related downregulation of miRNA expression is likely to represent an attempt to defend the respiratory tract from the adverse effects of CS, for instance by triggering antioxidant mechanisms, detoxification of carcinogens, DNA repair, antiinflammatory pathways, apoptosis, etc. On the other hand, a long-lasting exposure to CS can result in miRNA alterations triggering the activation of carcinogenic mechanisms, such as modulation of oncogenes and oncosuppressor genes, stimulation of cell proliferation, recruitment of undifferentiated stem cells, inflammation, inhibition of gap junctional intercellular communications, angiogenesis, invasion, and metastasis. As shown by our experimental data, the delicate balance between defense and adverse mechanisms renders CS a weak carcinogen after a short exposure, while the carcinogenic potential of CS can be expressed after a long-lasting exposure that results in persisting miRNA alterations. Furthermore, the carcinogenic response in mice is amplified when exposure to CS starts soon after birth,47,49 when the lung is particularly susceptible to genomic, transcriptional, and miRNA alterations.20,33,48

Our studies in rodents provided evidence that the CS-related miRNA dysregulation can be modulated by a variety of chemopreventive agents, of either dietary or pharmacological nature, that work with different mechanisms. As evaluated by testing NAC in ECS-exposed rats, modulation of miRNA expression correlates with modulation of protein expression. Our experimental approach is to investigate in parallel the effects of chemopreventive agents on miRNA expression in both CS-free and CS-exposed rodents. In this way, it is possible to evaluate whether the test agents do not appreciably affect the baseline expression of miRNAs, which is a molecular indicator of their potential safety. On the other hand, it is possible to ascertain whether the test agents are able to prevent CS-induced alterations of miRNA expression, which is a molecular indicator of their potential efficacy.22

The miRNA alterations induced by CS in the lung of mice and rats are similar to those observed in the airway epithelium of smoking humans,29 which indicates that the miRNA machinery is well conserved among species. Furthermore, the fact that several miRNAs that are modulated by either CS and/or chemopreventive agents are subjected to single nucleotide polymorphisms in humans53 is expected to help in predicting the individual response to both CS and chemopreventive agents according to toxicogenomic/pharmacogenomic approaches. Since miRNAs are released from target organs into the blood,54 thus providing a new biomarker in cancer research,55 future research will evaluate whether the analysis of blood miRNAs may provide an indicator of miRNA dysregulation by CS as well as safety/efficacy of chemopreventive agents in human trials by means of minimally invasive procedures.