Lung genotoxicity of benzo(a)pyrene in vivo involves reactivation of LINE-1 retrotransposon and early reprogramming of oncogenic regulatory networks

Abstract

Several lines of evidence have implicated long interspersed nuclear element-1 (LINE-1) retroelement in the onset and progression of lung cancer. Retrotransposition-dependent mechanisms leading to DNA mobilization give rise to insertion mutations and DNA deletions, whereas retrotransposition-independent mechanisms disrupt epithelial programming and differentiation. Previous work by our group established that tobacco carcinogens such as benzo(a)pyrene (BaP) reactivate LINE-1 in bronchial epithelial cells through displacement of nucleosome remodeling and deacetylase (NuRD) corepressor complexes and interference with retinoblastoma-regulated epigenetic signaling. Whether LINE-1 in coordination with other genes within its regulatory network contributes to the in vivo genotoxic response to BaP remains largely unknown. Evidence is presented here that intratracheal instillation of ORFeus^LSL mice with BaP alone or in combination with adenovirus (adeno)-CRE recombinase is genotoxic to the lung and associated with activation of the human LINE-1 transgene present in these mice. LINE-1 reactivation modulated the expression of genes involved in oncogenic signaling, and these responses were most pronounced in female mice compared with males and synergized by adeno-CRE recombinase. This is the first report linking LINE-1 and genes within its oncogenic regulatory network with early sexually dimorphic responses of the lung in vivo.

INTRODUCTION

Long interspersed nuclear elements (LINEs) are a group of non-long terminal repeat (non-LTR) retrotransposons widely distributed within the genome of eukaryotes (25, 51). LINEs account for ~20% of the human genome sequence (30), with most retroelements identified as molecular fossils devoid of activity (19, 30). About 100 full-length copies of human LINE-1 remain retrotransposition competent and capable of mobilization through RNA intermediates (4, 9). Full-length LINE-1 is ~6 kb in length, contains an internal bidirectional promoter, two open reading frames (ORFs) encoding ORF1 and ORF2, and a 3′ poly(A) tail. ORF1 is a 40-kDa protein with RNA-binding activity, whereas ORF2 is a 150-kDa protein with reverse transcriptase and endonuclease activities (16, 35). Both proteins are required for retrotransposition (35). The LINE-1 retrotransposition cycle entails three critical steps: 1) epigenetic reactivation and transcription of LINE-1, 2) cytoplasmic export of the mRNA for translation into proteins that bind the LINE-1 (L1) transcript to form ribonucleoprotein complexes that return to the nucleus, and 3) cleavage of genomic DNA by ORF2p to create a free 3′-end for reverse transcription as well as new sites for genome integration (43).

The majority of LINEs are silenced by DNA methylation and histone covalent modifications (11, 38, 55). Although human LINE-1 and mouse LINE-1 exhibit structural differences, benzo(a)pyrene (BaP) mediates transcriptional activation via comparable mechanisms involving disruption of epigenetic silencing, interference with retinoblastoma-regulated macromolecular interactions, and downregulation of DNA (cytosine-5)-methyltransferase 1 (DNMT1) expression (38, 55). Several lines of evidence in vitro established the ability of LINE-1 to retrotranspose to new genomic locations where it compromises genome stability through altered gene splicing, disruption of gene function, and increased recombination (7, 15, 53). LINEs may also control gene expression by providing regulatory sequences that direct expression of other genes (24, 56). The activity of LINE-1 in vivo has not been widely studied, and to date, only two active human LINEs, L1RP (10) and L1LRE3 (13), have been described. These LINEs exhibit extremely low retrotransposition frequencies when introduced into mice in their native form, and genetic manipulation is required to optimize expression (2).

BaP is a by-product of the incomplete combustion of organic matter (47) and the principal constituent of tobacco smoke implicated in lung carcinogenesis (28). The parent hydrocarbon and its oxidative metabolites bind to the aryl hydrocarbon receptor (AhR) to induce nuclear translocation and dimerization with the AhR nuclear translocator to form a liganded complex that regulates several genes, including members of the cytochrome P450 (CYP) superfamily (36). The relative abundance of CYP and AhR proteins in the lung is high, with CYPs readily catalyzing the conversion of BaP to (±)-anti-benzo(a)pyrene-7,8-diol-9,10-epoxide (BPDE) and other metabolites that form the DNA adducts detected in the lung tissue of smokers (40). BaP activates LINE-1 expression in various cell types and multiple mammalian species (33, 34, 54), and AhR signaling participates in the LINE-1 reactivation cascade (54).

The present study evaluated profiles of LINE-1 reactivation and genetic reprogramming following genotoxic lung injury by BaP. The ORFeus^LSL murine transgenic model containing a single copy of an optimized human LINE-1 transgene was employed to define exogenous activation. We report that the genotoxicity of BaP alone or in combination with adenovirus (adeno)-CRE recombinase was associated with activation of the LINE-1 transgene and a sexually dimorphic profile of genetic reprogramming linked to early oncogenic signaling.

MATERIALS AND METHODS

Animals.

Founder ORFeus^LSL mice, a genetically modified conditional model of LINE-1 retrotransposition under control of CRE recombinase, were provided by Dr. W. An (South Dakota State University). All procedures were reviewed and approved by the University of Arizona Institutional Animal Care and Use Committee.

Reactivation of ORFeus^LSL transgene by BaP alone or with adeno-CRE recombinase.

BaP dissolved in dimethyl sulfoxide (DMSO) was administered via instillation with or without adeno-CRE recombinase for conditional activation of a single copy of a human LINE-1 transgene (2). A BioLite Intubation System (Braintree Scientific) was used for intratracheal instillation of ORFeus^LSL mice (29–38 g) anesthetized with ketamine and xylazine (100 and 10 mg/kg, respectively). All mice were genotyped to confirm genetic identity. Control mice were given 1 dose of 5 × 10¹⁰ plaque-forming units (PFU) adeno-CRE recombinase dissolved in H₂O or DMSO. The optimal adenovirus dose was defined empirically as a dose that only modestly activated the transgene (data not shown). Treated mice were intubated with BaP (50 mg/kg) alone or in combination with 5 × 10¹⁰ PFU adeno-CRE virus. Cervical dislocation of mice was performed after euthanasia with CO₂ 1 wk after treatment. Tissues were collected and stored at 4°C in RNA later (Thermo Fisher Scientific).

Histological analysis.

Lung sections were snap-frozen in optimum cutting temperature compound. Tissues were fixed with cold acetone and stained with hematoxylin-eosin. Samples were evaluated by two pathologists blinded to the treatment given.

³²P postlabeling.

The ³²P-postlabeling assay for DNA adducts was performed as reported previously (39).

Quantitative real-time PCR.

RNA was extracted using the RNeasy Plus Mini Kit (Qiagen). An aliquot of 250 ng RNA was reverse transcribed into first-strand cDNA using SuperScript III (Invitrogen). Quantitative real-time PCR (RT-qPCR) was performed for LINE-1 and nine genes within its oncogenic regulatory network. These genes included AhR; chemokine (C-C motif) ligand 2 (CCL2); cytochrome P450, family 2, subfamily a, polypeptide 4 (CYP2A4); microsomal glutathione S-transferase 1 (MGST1); phenylalanine hydroxylase (PAH); periostin, osteoblast-specific factor (POSTN); protein tyrosine phosphatase, receptor type B (PTPRB); vascular cell adhesion molecule 1 (VCAM1); and transforming growth factor-β1 (TGF-β1; 8, 45). GAPDH was used as a control. Each reaction included 4.0 μL of cDNA, 0.5 μL of gene-specific primer pairs, and 5.0 μL SYBR Green dye, and double-distilled H₂O to a final volume of 20 μL. Amplifications were completed using StepOnePlus RT-qPCR System and fold changes were calculated using the 2^−∆∆Ct method (where Ct is threshold cycle).

Statistical analyses.

All normally distributed data were analyzed by ANOVA and Duncan’s multiple range test. Nonnormally distributed data were evaluated using the Mann–Whitney U test. Significance was defined at the P ≤ 0.05 level.

RESULTS

BPDE-deoxyguanine (dG) DNA adducts were detected in the lungs of both female and male ORFeus^LSL mice given a single dose of 25 mg/kg BaP alone or in combination with adeno-CRE (Table 1). Only background signal was detected in control untreated, CRE-H₂O, or CRE-DMSO mice. All major adducts were derived from BPDE, as evidenced by autoradiographic signals (3, 6). Adduct intensities were considerably higher in female mice compared with males, with significant differences between BaP alone and BaP plus adeno-CRE seen only in male mice. The occurrence of BPDE-dG DNA adducts 1 wk following carcinogen treatment established persistence of covalent binding following widespread DNA damage. This is significant given that the BaP dose given is well below the total cumulative dose range of 25–100 mg/kg administered daily over several months for human toxicological risk assessments (52). No dramatic toxic effects were observed at any time in mice treated with BaP alone or in combination with adeno-CRE recombinase.

Table 1.

DNA adduct signal intensity in the lungs of ORFeus^LSL mice following treatment with benzo(a)pyrene alone or in combination with CRE recombinase

	Relative Adduct Labeling
Treatment	Adduct 1	Adduct 2	Adduct 3	Adduct 4	Adduct 5	Total
Untreated control	3.50 ± 1.91f	1.69 ± 0.88n,o	2.71 ± 1.65j	1.39 ± 0.62q	1.68 ± 0.34n,o	2.19 ± 0.82E
CRE-DMSO	2.82 ± 1.95i	1.71 ± 1.23n	1.92 ± 0.80m	0.84 ± 0.35u,v	1.18 ± 0.29s	1.69 ± 0.70F
BaP (F)	28.01 ± 17.49a	2.95 ± 0.99h	4.04 ± 1.74e	1.44 ± 0.65p,q	1.40 ± 0.83q	7.57 ± 10.92A
BaP (M)	10.35 ± 6.21d	1.65 ± 0.30o	1.48 ± 0.78p	0.66 ± 0.23w	1.25 ± 0.43r	3.08 ± 3.78D
Cre-BaP (F)	23.86 ± 3.99b	2.39 ± 0.87k	3.10 ± 0.66g	0.89 ± 0.16u	1.30 ± 0.55r	6.31 ± 9.12B
Cre-BaP (M)	17.61 ± 9.34c	1.94 ± 0.35m	2.29 ± 0.51l	0.80 ± 0.28v	1.05 ± 0.72t	4.73 ± 6.68C
Total	14.36 ± 9.89A	2.06 ± 0.48B	2.59 ± 0.86B	1.0 ± 0.30C	1.31 ± 0.22C

Values are means ± SD; n = 3 animals examined. Relative adduct labeling is equal to the number of adducts per 108 nucleotides. BaP, benzo(a)pyrene; F, female; M, male.

^{a–w, A–F}Significant differences between and among the different groups by sex × treatment.

Control lung tissues from mice given water or DMSO alone, or in combination with adeno-CRE recombinase, showed normal histology, except for a modest degree of iatrogenic pulmonary edema evidenced by vascular dilation (Fig. 1, A–E). In contrast, significant leukocytic infiltration into the lungs was prominent in mice given BaP alone or combined with adeno-CRE. Female and male mice in both treatment groups showed pulmonary vessel congestion with focal-to-diffuse mononuclear cell infiltration into parenchyma. Lung tissues also showed emphysematous changes evidenced by large empty spaces with variable preservation of alveolar structures (Fig. 1, F–I).

Fig. 1.

Histology of lung tissues in ORFeus^LSL mice treated with benzo(a)pyrene (BaP) alone or in combination with CRE recombinase. A–I: representative results for all treatment groups at ×20. Blue arrows indicate the pulmonary edema evidenced by vascular dilation, black arrows denote pulmonary vessel congestion, and the red arrow denotes emphysematous changes evidenced by large empty spaces with variable preservation of alveolar structures. Here, n = 3 animals examined.

Measurements of ORFeus LINE-1 mRNA were completed to evaluate the ability of the carcinogen to activate the human transgene. A schematic of the ORFeus^LSL transgene under control by loxP is presented in Fig. 2A. BaP alone significantly increased expression of the transgene, whereas adeno-CRE recombinase alone was only modestly active. Increased expression of ORFeus LINE-1 mRNA was observed in both male and female mouse lungs given the combined BaP-adeno-CRE treatment compared with BaP alone or vehicles. In keeping with patterns of DNA adduct formation and tissue damage, ORFeus LINE-1 mRNA levels were higher in female mouse lungs compared with male mouse lungs (Fig. 2B). The specificity of the lung induction response was confirmed in studies showing that only stomach, but not heart, kidney, liver, or intestine, showed activation of the LINE-1 transgene and that BaP treatment induced CYP gene expression in the lungs of ORFeus mice (data not shown).

Fig. 2.

A: schematic of the ORFeus^LSL transgene: 1) the floxed allele of the ORFeus^LSL transgene, 2) the excised allele of the ORFeus^LSL transgene following Cre-mediated excision of the floxed β-geo-stop cassette (LSL), and 3) putative insertion upon retrotransposition. B: long interspersed nuclear element-1 (LINE-1) mRNA in lungs of ORFeus^LSL mice treated with benzo(a)pyrene (BaP) alone or in combination with CRE recombinase. LINE-1 mRNA was measured by RT-qPCR using primers specific for ORFeus^LSL. A (upright), polyadenylation signal; gfp, green fluorescent protein; LTR, long terminal repeat; ORF, open reading frame; RQ, relative quantification; n = 3 animals examined. Different letters denote differences between control, BaP-treated, and CRE-BaP-treated groups at P ≤ 0.05.

To investigate the biological consequences of LINE-1 activation in the lungs of ORFeus mice, we next examined genes within a LINE-1 regulatory network that is activated during oncogenic reprogramming of epithelial genomes (8, 45). The targets chosen constitute a regulatory network involved in early oncogenic signaling and included AhR, CCL2, CYP2A4, MGST1, PAH, POSTN, PTPRB, VCAM1, and TGF-β1. These genes were differentially regulated upon activation of the LINE-1 transgene, with notable differences between BaP alone or in combination with adeno-CRE seen for all but POSTN and PTPRB (Fig. 3, A–I). All mice were of comparable age to obviate potential age-related differences. Thus, induction of ORFeus LINE-1 was associated with genetic reprogramming in the murine lung in vivo, and this response was sustained for up to 1 wk after carcinogen treatment.

Fig. 3.

mRNA abundance of genes within the long interspersed nuclear element-1 oncogenic regulatory network in lungs of ORFeus^LSL mice. A–I: results for different genes within the network: AhR, aryl hydrocarbon receptor; CCL2, chemokine (C-C motif) ligand 2; CYP2A4, cytochrome P450, family 2, subfamily a, polypeptide 4; MGST, microsomal glutathione S-transferase 1; PAH, phenylalanine hydroxylase; POSTN, periostin, osteoblast-specific factor; PTPRB, protein tyrosine phosphatase, receptor type B; TGFB1, transforming growth factor-β1; VCAM, vascular cell adhesion molecule 1. BaP, benzo(a)pyrene; RQ, relative quantification; n = 3 animals examined. Different letters denote differences between control, BaP-treated, and CRE-BaP-treated groups at P ≤ 0.05.

DISCUSSION

BaP is a procarcinogen metabolized by CYP1A1 and CYP1B1 to highly reactive electrophilic intermediates (17, 26). This metabolic activation mediates DNA adduct formation, including those associated with BPDE, the primary lung cancer etiologic agent (27, 50). The total cumulative dose of BaP examined in our study approximates the cumulative dose experienced by a smoker consuming 40 cigarettes per day for 2 yr. This is significant given that 1 yr in the life of a human is equivalent to 9 days in the mouse (14). As such, our findings reflect early changes associated with activation of the LINE-1 retrotransposon and genetic elements within its oncogenic regulatory network.

BaP is rapidly distributed throughout the body and detected in most tissues within minutes to hours after exposure. DNA adduction, a key event in the mutagenic and carcinogenic response, appears as early as 4 h after dosing of mice with the parent hydrocarbon (18). We present evidence for the first time that the genotoxic response of the mouse lung to BaP in vivo involves sustained activation of LINE-1 retrotransposon and this response exhibits a sexually dimorphic profile. These findings impact our present understanding of tobacco-induced carcinogenesis.

Sex differences in BaP metabolism and DNA repair capacity have been documented (49), with several studies indicating that relative lung cancer risk among females exceeds that of males by 2.5-fold (21, 48, 60). In our study, the levels of BaP-DNA adducts were higher in the lungs of female mice compared with males, a finding consistent with the enhanced female susceptibility to tobacco carcinogens and sex-related differences in CYP gene expression between women and men (20, 37). Increased lung cancer susceptibility in females may also be strongly influenced by interactions between estrogen receptors and proteins involved in the regulation of hydrocarbon metabolism (16). Estrogen receptors are expressed in normal lung tissue as well as tumors (5) and strongly implicated in lung development (22).

In most human and rodent somatic cells, LINEs are silenced epigenetically via DNA methylation, a process that functions in tandem with LINE-1 deamination by DNA dC->dU-editing enzyme APOBEC (APOBEC) proteins, degradation of LINE-1 mRNA by three-prime repair exonuclease 1 (TREX1) and Piwi-interacting RNAs, and transcriptional repression by epigenetic modifying proteins (29). As such, activation of a LINE-1 transgene engineered to operate under the regulatory control of CRE recombinase in vivo provides compelling evidence that BaP overrides regulatory control to persistently activate LINE-1 transcription and to activate its genetic regulatory network. Given the potentially devastating consequences of LINE-1 reactivation, these findings suggest that LINE-1 and genes within its oncogenic regulatory network play a major role in the lung pathologic response to tobacco carcinogens. Although the degree to which transcriptional activation of LINE-1 and/or retrotransposition couples with DNA damage in vivo to modulate carcinogenesis requires further investigation, LINE-1 is associated with insertion mutations, genetic deletions, and reprogramming of epithelial differentiation (29). As such, the observed responses in vivo raise important questions about the contributions of retrotransposition-dependent and -independent mechanisms to BaP carcinogenesis in vivo. In accord, oncogenic transformation is associated with elevated LINE-1 expression in human lung tumors in the absence of changes in neighboring tissues (46).

Previous work by our group discretized a LINE-1 genetic regulatory network that is linked to host regulation of LINE-1 and oncogenic signaling (8, 45). In vivo validation of the functional integrity of this regulatory network in the lungs of ORFeus mice indicates that genes within the network are involved in the acute and adaptive responses of the lung to BaP. For instance, AhR-dependent genetic regulation of lung tissue has been associated with inflammation, DNA adduct formation, epithelial-mesenchymal transition, and tumorigenesis (23, 31, 57). The diffuse infiltration of mononuclear cells in BaP-treated lung tissue is consistent with the upregulation of CCL2, a major chemoattractant involved in leukocyte homing to sites of inflammation (32). The upregulated expression of MGST1 and CYP2A4 is often secondary to disruption of cellular defenses by electrophilic compounds through conjugation of reduced glutathione and compensatory feedback control (42). BaP induces genes involved in epithelial reprogramming such as POSTN (59) and PTPRB, the latter being a protein tyrosine phosphatase involved in oncogenic transformation (12). VCAM1, a member of the immunoglobulin superfamily (61), is elevated in idiopathic pulmonary fibrosis by TGF-β1 (1), a known inducer of epithelial dedifferentiation (44). Thus, our findings advance our understanding of the role of LINE-1 during the early stages of oncogenic signaling activated by tobacco lung carcinogens in vivo and provide a platform for evaluation of treatment modalities for precision management of thoracic malignancies.

GRANTS

Funding for this work was provided in part by University of Arizona Health Sciences and the Texas Governor’s University Research Initiative to K. S. Ramos, an Egyptian Science and Technology Development Fund Junior Scientist Development visit grant, Cycle 17, to A. A. I. Hassanin, and NIH Grants R01 ES-029382 and R01 HL-129794 to B. Moorthy.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.A.I.H., M.T.-G., B.M., and K.S.R. conceived and designed research; A.A.I.H., M.T.-G., and G.-D.Z. performed experiments; A.A.I.H., B.M., G.-D.Z., and K.S.R. analyzed data; A.A.I.H., M.T.-G., B.M., G.-D.Z., and K.S.R. interpreted results of experiments; A.A.I.H. prepared figures; A.A.I.H. drafted manuscript; A.A.I.H., M.T.-G., B.M., and K.S.R. edited and revised manuscript; K.S.R. approved final version of manuscript.