Chronic Inhalation Exposure to Antimony Trioxide Exacerbates the MAPK Signaling in Alveolar Bronchiolar Carcinomas in B6C3F1/N Mice

Abstract

Antimony trioxide (AT) is used as a flame retardant in fabrics and plastics. Occupational exposure in miners and smelters is mainly through inhalation and dermal contact. Chronic inhalation exposure to AT particulates in B6C3F1/N mice and Wistar Han rats resulted in increased incidences and tumor multiplicities of alveolar/bronchiolar carcinomas (ABCs). In this study, we demonstrated Kras (43%) and Egfr (46%) hotspot mutations in mouse lung tumors (n = 80) and only Egfr (50%) mutations in rat lung tumors (n = 26). Interestingly, there were no differences in the incidences of these mutations in ABCs from rats and mice at exposure concentrations that did and did not exceed the pulmonary overload threshold. There was increased expression of p44/42 mitogen-activated protein kinase (MAPK) (Erk1/2) protein in ABCs harboring mutations in Kras and/or Egfr, confirming the activation of MAPK signaling. Transcriptomic analysis indicated significant alterations in MAPK signaling such as ephrin receptor signaling and signaling by Rho-family GTPases in AT-exposed ABCs. In addition, there was significant overlap between transcriptomic data from mouse ABCs due to AT exposure and human pulmonary adenocarcinoma data. Collectively, these data suggest chronic AT exposure exacerbates MAPK signaling in ABCs and, thus, may be translationally relevant to human lung cancers.

Keywords

antimony trioxide alveolar/bronchiolar carcinoma immunohistochemistry inhalation exposure MAPK pathway mouse

Introduction

Antimony trioxide (AT), or Sb₂O₃, is used as a flame retardant in textiles, paper, and plastics. AT is also used in combination with some chlorinated or brominated flame retardants on commercial furniture, draperies, wall coverings, and carpets.^1,2 Workers are exposed to antimony in the metal ore smelting and mining industries mainly through inhalation and dermal contact with dusts.³ In the National Toxicology Program’s rodent cancer bioassay,⁴ chronic inhalation exposure of Wistar Han rats and B6C3F1/N mice to AT particulates for 2 years resulted in increased incidences of alveolar/bronchiolar adenomas (ABAs) and carcinomas (ABCs) (Figure 1, Tables 1 and 2).

Figure 1.

Alveolar/bronchiolar carcinoma, hematoxylin and eosin. A solid, densely cellular mass in the lung of a female B6C3F1/N mouse exposed to 30 mg of antimony trioxide/m³ by inhalation for 2 years. Please note the scattered black aggregates of antimony trioxide dust within the parenchyma. To access the whole-slide image provided with this figure, scan the QR code on cover page 3 or visit slide 1 .

Table 1.

Incidences of alveolar/bronchiolar tumors in Wistar Han rats after 2-year inhalation exposure to antimony trioxide (NTP TR 590).

Tumor type	Chamber control	3 mg/m³	10 mg/m³	30 mg/m³
Males
Alveolar/bronchiolar adenomas, multiple	0	1	0	3
Alveolar/bronchiolar adenomas (including multiples)^a,b	3/50 (6%)	4/50 (8%)	6/50 (12%)	8/50 (16%)
Poly-3 test*	P = .057	P = .478	P = .253	P = .083
Alveolar/bronchiolar carcinomas^c	0/50 (0%)	0/50 (0%)	2/50 (4%)	0/50 (0%)
Poly-3 test*	P = .702 N	—	P = .245	—
Alveolar/bronchiolar adenomas and carcinomas (combined)^d	3/50 (6%)	4/50 (8%)	8/50 (16%)	8/50 (16%)
Poly-3 test*	P = .067	P = .478	P = .104	P = 0.083
Females
Alveolar/bronchiolar adenomas, multiple	0	0	1	0
Alveolar/bronchiolar adenomas (including multiples)^e	0/50 (0%)	2/50 (50%)	6/50 (12%)	5/50 (10%)
Poly-3 test*	P = .029	P = .235	P = .012	P = .021
Alveolar/bronchiolar carcinomas	0/50 (0%)	0/50 (0%)	0/50 (0%)	0/50 (0%)
Alveolar/bronchiolar adenomas and carcinomas (combined)	0/50 (0%)	2/50 (4%)	6/50 (12%)	5/50 (10%)
Poly-3 test*	P = .029	P = .235	P = .012	P = .021

Table #9 from the NTP TR on the toxicology and carcinogenesis studies of antimony trioxide (TR 590).

Abbreviations: NTP, national toxicology program; TR, technical report.

Historical incidence for 2-year inhalation studies with chamber control groups (mean ± standard deviation): 4/150.

Number of animals with neoplasm per number of animals with lung examined microscopically.

Historical incidence for inhalation studies: 0/150; all routes: 0/299.

Historical incidence for 2-year inhalation studies with chamber control groups (mean ± standard deviation): 4/150 (2.7% ± 3.1%), range 0% to 6%; all routes: 4/299 (1.3% ± 2.4%), range 0% to 6%.

Historical incidence for inhalation studies: 0/150; all routes: 0/300.

P value listed under chamber control is associated with the trend test. P values listed under each exposed group correspond to pairwise comparisons between the chamber controls and the exposed group. The poly-3 test accounts for differential mortality in animals that do not reach terminal kill. A negative trend is indicated by N.

Table 2.

Incidences of alveolar/bronchiolar tumors in B6C3F1/N mice after 2-year inhalation exposure to antimony trioxide (NTP TR 590).

Tumor type	Chamber control	3 mg/m³	10 mg/m³	30 mg/m³
Males
Alveolar/bronchiolar adenomas, multiple	3	3	2	4
Alveolar/bronchiolar adenomas (including multiples)^a	10/50 (20%)	14/50 (28%)	9/50 (18%)	14/50 (28%)
Poly-3 test^b	P = .190	P = .165	P = .592	P = .129
Alveolar/bronchiolar carcinomas, multiple	0	5*	6**	11**
Alveolar/bronchiolar carcinomas (including multiples)^c	4/50 (8%)	18/50 (36%)	20/50 (40%)	27/50 (54%)
Poly-3 test	P < .001	P < .001	P < .001	P < .001
Alveolar/bronchiolar adenomas or carcinomas (combined)^d	13/50 (26%)	29/50 (58%)	28/50 (56%)	34/50 (68%)
Poly-3 test	P < .001	P < .001	P < .001	P < .001
Females
Alveolar/bronchiolar adenomas, multiple	0	3	3	1
Alveolar/bronchiolar adenomas (including multiples)^e	1/50 (2%)	10/50 (20%)	19/50 (38%)	8/50 (16%)
Poly-3 test*	P = .137	P = .003	P < .001	P = .009
Alveolar/bronchiolar carcinomas, multiples	0	7**	6*	4*
	0/50 (0%)	0/50 (0%)	0/50 (0%)	0/50 (0%)
Alveolar/bronchiolar carcinomas (including multiples)^f	2/50 (4%)	14/50 (28%)	11/50 (22%)	11/50 (22%)
Poly-3 test	P = .065	P < .001	P = .003	P = .002
Alveolar/bronchiolar adenomas or carcinomas (combined)^g	3/50 (6%)	22/50 (44%)	27/50 (54%)	18/50 (36%)
Poly-3 test	P = .019	P < .001	P < .001	P < .001

Table #22 from the NTP TR on the toxicology and carcinogenesis studies of antimony trioxide (TR 590).

Abbreviations: NTP, national toxicology program; TR, technical report.

Historical incidence for 2-year inhalation studies with chamber control groups (mean ± standard deviation): 35/250 (14.0% ± 3.7%), range 10% to 20%; all routes: 83/550 (15.1% ± 5.9%), range 8% to 26%.

Beneath the chamber control incidence is the P value associated with the trend test. Beneath the exposed group incidence are the P values corresponding to pairwise comparisons between the chamber controls and that exposed group. The poly-3 test accounts for differential mortality in animals that do not reach terminal kill.

Historical incidence for inhalation studies: 42/250 (16.8% ± 5.4%), range 8% to 22%; all routes: 75/550 (13.6% ± 6.4%), range 4% to 22%.

Historical incidence for inhalation studies: 69/250 (27.6% ± 2.6%), range 26% to 32%; all routes: 147/550 (26.7% ± 6.5%), range 16% to 38%.

Historical incidence for inhalation studies: 12/249 (4.8% ± 2.7%), range 2% to 8%; all routes: 27/549 (4.9% ± 3.5%), range 0% to 10%.

Historical incidence for inhalation studies: 17/249 (6.8% ± 3.7%), range 2% to 10%; all routes: 24/549 (4.4% ± 3.5%), range 0% to 10%.

Historical incidence for inhalation studies: 28/249 (11.3% ± 5.5%), range 6% to 18%; all routes: 50/549 (9.1% ± 5.2%), range 2% to 18%.

Significantly different (P ≤ .05) from the chamber control group by the Fisher exact test (interim evaluation) or the poly-3 test (2-year study).

P ≤ .01.

Somatic mutations in oncogenes such as Kras and Egfr (both major components of the MAP kinase signaling pathway) are common driver mutations in human lung cancers.^5
-8 In rodent lung tumors, these mutations may arise either spontaneously or due to chronic chemical exposures. The mutation spectra observed in rodent tumors arising from certain chronic chemical exposures are often conserved across species and are also observed in certain human cohorts. For example, G to T transversions in codon 12 of Kras gene are commonly noted in lung cancers resulting from tobacco smoke exposure and also in rodent lung tumors resulting from chronic exposure to cobalt, which is mechanistically attributed to metal-induced oxidative stress.^9,10

In human non-small-cell lung cancer (NSCLC), the incidence of Kras mutations is approximately 26%, with the majority of the point mutations seen in codon 12 followed by fewer mutations in codons 13 and 61.^5,11 Point mutations in codons 12, 13, and 61 of Kras are activating mutations that result in constitutive activation of the KRAS protein, making it refractory to the inhibitory GTPase-activating proteins, which results in stimulus-independent, persistent activation of downstream effectors, in particular, the RAF-MEK-ERK cascade. Constitutive activation of this kinase cascade results in cellular proliferation and transformation.^6,7

The overall incidence of Egfr mutations in human NSCLC, a subtype of adenocarcinoma, differs by region with the highest in the Asia-Pacific region (47%) and lowest in the Oceania region (12%).¹² The majority (70%) of Egfr mutations are observed within exons 19 and 21.⁵ Epidermal growth factor receptor (EGFR) is a transmembrane receptor, which upon ligand binding, dimerizes and activates the cytosolic kinase domain of the receptor tyrosine kinase, resulting in activation of signaling pathways, such as the aforementioned mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) pathway, that support cancer development and progression. Other signaling pathways that can be activated include the phosphatidylinositol 3-kinase (PI3K) pathway, which, when activated, leads to AKT activation and apoptosis inhibition.

In the current study, we have examined the hotspot mutations in Kras and Egfr (rats and mice) and transcriptomic alterations (mice only) in ABCs resulting from chronic inhalation exposure to AT for 2 years to better understand the translational mechanistic relevance of these rodent tumors to human lung cancers.

Materials and Methods

Animal Tissue Samples

Frozen lung samples and formalin-fixed, paraffin-embedded (FFPE) lung tissues were obtained from male and female Wistar Han rats (Charles River Laboratories, Raleigh, North Carolina) and from male and female B6C3F1/N mice (Taconic Farms, Germantown, New York) used in the AT cancer bioassay. The lung tumors were diagnosed ABAs and ABCs based on microscopic morphologic criteria described in the INHAND (International Harmonization of Nomenclature and Diagnostic Criteria) document on rat and mouse respiratory system. The diagnostic terminology “alveolar/bronchiolar” is the preferred NTP terminology and is slightly different from the “bronchiolo-alveolar” terminology described in the INHAND document. All the diagnoses were confirmed by an NTP pathology working group. The ABAs and ABCs that arose either spontaneously in chamber controls or due to chronic inhalation exposure to AT had similar morphologic features with the exception of increased tumor multiplicity and the test material accumulation in the exposed lungs. Animal care and use were in accordance with the Public Health Service Policy on Humane Care and Use of Animals. Animal studies were in compliance with the IACUC (Institutional Animal Care and Use Committee) protocol approved by the Battelle Toxicology Northwest Animal Care and Use Committee and were conducted in an animal facility accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care International.

Alveolar/Bronchiolar Neoplasms for Mutation Analysis

Morphologic evaluation based on the INHAND criteria was performed to diagnose and select alveolar/bronchiolar neoplasms. Alveolar/bronchiolar adenomas (n = 23) and ABCs (n = 3) from Wistar Han rats and ABCs (n = 80) from B6C3F1/N mice were used from AT-exposed groups. Rat ABAs (n = 2) and mouse ABCs (n = 9) obtained from the chamber controls were considered to have arisen spontaneously. Nontumor lung tissues from chamber control rats (n = 11) and mice (n = 10) were used as negative controls. A mutation analysis was performed on FFPE tissues. The size of the selected mouse tumors was generally greater than 5 mm in diameter. Rat tumors were smaller than 5 mm in diameter and scattered throughout the pulmonary parenchyma. The alveolar/bronchiolar tumors selected for molecular biology analysis were based on their overall size and tissue viability (minimal to no necrosis, autolysis, or hemorrhage microscopically) in order to maximize the amount and quality of DNA obtained from the FFPE sections. DNA quantity was measured on a NanoDrop spectrophotometer (Thermo Fisher Scientific, Wilmington, Delaware). DNA sample quality was estimated by 260/280-nm absorbance ratio; samples with an absorbance ratio range between 1.7 and 2.0 were used for analysis. Samples outside this DNA purity range were re-isolated from FFPE sections until a suitable purity measure was obtained.

DNA Extraction, Polymerase Chain Reaction, Autosequencing, and Mutation Analysis

The lung tumors were evaluated for hot spot mutations within specific codons or exons in Kras and Egfr genes that are relevant in human lung cancer. Ten-micron-thick FFPE sections were collected on charged glass slides, and the tumor-only tissue was grossly dissected using a sharp microtome blade and collected into screw-top tubes. In some cases (especially in rats) where the alveolar/bronchiolar tumors had a miliary distribution throughout the pulmonary parenchyma, the entire lung section was used for DNA isolation. DNA was isolated from these FFPE sections using a DNeasy Tissue Kit (Qiagen, Valencia, California) following the manufacturer’s protocols. Amplification reactions were carried out by semi-nested polymerase chain reaction (PCR) using primer sets designed for Kras (exons 1 and 2) and Egfr (exons 18 to 21).⁹ Controls without template DNA were run with all sets of reactions. PCR products were purified using a QIAquick Gel Extraction Kit (Qiagen). The purified products were cycled with Terminal Ready Reaction Mix-Big Dye (PerkinElmer, Foster City, California), and the extension products were purified with the DyeEx 2.0 SpinKit (Qiagen). The lyophilized PCR products were sequenced with an automatic sequencer (PerkinElmer ABI Model 3100).¹³ The resulting electropherograms were compared to identify mutations in alveolar/bronchiolar neoplasms that either arose spontaneously or were due to AT exposure using CLC Genomic Workbench (Redwood City, California).

Statistics for Mutation Analysis

A one-sided Cochran-Armitage trend test was conducted to test for significance of exposure concentration-related trends in the incidences of mutations. One-sided Fisher’s exact tests were conducted to test for significant differences in the proportions of mutations between the chamber control and various exposed groups.

RNA Isolation

Mouse ABCs arising either spontaneously (n = 6) in chamber controls or due to chronic AT exposure (n = 6) were used for transcriptomic analysis. RNA was extracted from tumor-only tissue obtained using laser-capture microdissection of the frozen mouse ABCs collected from the AT cancer bioassay. RNA from mouse nontumor lung tissue (n = 6) was used as negative controls. Invitrogen PureLink Mini kit (Invitrogen, Carlsbad, California) was used for RNA extraction following manufacturer’s protocol. On-column deoxyribonuclease (DNase) treatment was performed using the Invitrogen PureLink DNase kit (Invitrogen) to purify RNA samples. The concentration and quality of the extracted RNA were measured on a Bioanalyzer (Agilent Technologies, Santa Clara, California). Samples were aliquoted and stored at −80°C until they were analyzed. All the RNA samples used for the microarray study had an RNA integrity number value of at least 7.

Microarray Protocol

A global differential gene expression analysis was done using the Affymetrix Mouse Genome 430 2.0 GeneChip arrays (Affymetrix, Santa Clara, California). Twenty nanograms of total RNA was amplified as directed in the WT-Ovation Pico RNA Amplification System (Nugen, San Carlos, California) protocol and labeled with biotin following the Encore Biotin Module. Amplified biotin-aRNAs (4.6 μg) were fragmented and hybridized to each array for 18 hours at 45°C by rotating hybridization. Array slides were stained with streptavidin/phycoerythrin utilizing a double-antibody staining procedure and then washed for antibody amplification according to the GeneChip Hybridization, Wash, and Stain Kit user manual following protocol FS450-0004. Arrays were scanned in an Affymetrix Scanner 3000, and data were obtained using the GeneChip Command Console and Expression Console Software (AGCC, Version 3.2, and Expression Console, Version 1.2) using the MAS5 algorithm to generate .CHP files. Preliminary analyses were performed with OmicSoft Array Studio (Version 7.0) software (Qiagen, Germantown, MD).

Microarray Data Analysis

Gene expression data were normalized across all samples using the robust multiarray analysis (RMA) methodology¹⁴ in the R statistical software environment. RMA analysis fits a linear additive model to the log-transformed perfect match values using median polish to produce the gene expression measures used in the statistical analyses. RMA-normalized data were used for statistical analysis to determine differentially expressed genes. For each probe set, a pairwise t-test was performed between groups of samples (either control vs spontaneous ABC or control vs AT-exposed ABC). In each case, the null distribution of the t-statistic was derived using 10,000 bootstrap samples obtained by resampling the residuals in the ORIOGEN (Shyamal Peddada at NIEHS, https://www.niehs.nih.gov/research/resources/software/biostatistics/oriogen/index.cfm.) software package.¹⁵ A false discovery rate (FDR) threshold of 0.05 was used to establish statistical significance.

Partek Genomics Suite, version 7.0 (release date November 25, 2019; Partek, St. Louis, Missouri), was used to perform principal components (PC) analysis (PCA) on the complete catalog of probe sets from the normalized gene expression data and to generate heat maps to compare the control lung samples, spontaneous ABC tumor samples, and AT-exposed ABC tumor samples for differentially expressed probe sets. PCA uses linear transformation to reduce the dimension of the data from p variables to k PCs. The first three PCs capture about 50% of the variation in the data and were used to visualize the spatial relationship of the control, spontaneous ABC, and AT-exposed ABC tumor samples. For hierarchical clustering, gene expression measures were first standardized across samples in a probe-set-specific manner in which the mean expression level of a transcript is subtracted from each gene expression value and the quantity is divided by the standard deviation of values across all samples. After standardization, the probe sets (rows) and samples (columns) of the data matrix holding gene expression measures for significant genes were subjected to agglomerative hierarchical clustering, where Pearson’s dissimilarity was used as the distance measure and Ward’s method was used to evaluate the distance between clusters.¹⁶ Microarray data files (.cel) and associated annotations have been submitted to the GEO database: accession GSE152629 (reviewer token “wbubysqchpknnsd”).

Determination of Overrepresented Canonical Pathways

Ingenuity Pathway Analysis (IPA), application building_utopia, version 47547484, was used to evaluate the most statistically significant overrepresented or enriched canonical pathways included in the Ingenuity knowledge base. The significant biological canonical pathways were derived from the entire IPA library, and P < 0.01 (Fisher’s exact test) was used to indicate statistical significance.

Comparison of AT-Exposed ABCs from Mice to Human NSCLCs

Illumina BaseSpace Correlation Engine (previously known as NextBio; Illumina, Santa Clara, California) was used to compare the transcriptomic data sets of AT-exposed mouse ABCs from this study to human NSCLCs and identify the common pathways enriched in both species. Illumina BaseSpace Correlation Engine is a curated and correlated repository of experimental data derived from an extensive set of public sources (e.g., ArrayExpress and GEO) that allows the user to compare patterns of gene expression in their experiment to the published data sets. Statistical analysis is carried out using rank-based enrichment analysis to compute pairwise correlation scores of the uploaded data set and all studies in the Illumina BaseSpace Correlation Engine. The statistical analysis method used by the Illumina BaseSpace Correlation Engine is referred to as a “Running Fischer” algorithm and is very similar to the gene set enrichment analysis.¹⁷ A detailed explanation of the analysis protocol used by the Illumina BaseSpace Correlation Engine is described elsewhere.¹⁸

The Illumina BaseSpace Correlation Engine software was utilized to compare the differential transcriptomic changes common to human NSCLCs (curated by the Illumina BaseSpace Correlation Engine) and the AT-exposed mouse ABCs. The top-ranked upregulated and downregulated genes, as well as functional categories (termed biogroups by Illumina BaseSpace Correlation Engine) and canonical pathways that are common to AT-exposed mouse ABCs and human NSCLC, were compared. A biogroup’s score is based on the overall statistical significance and consistency of the enrichment, or overlap, between the set of genes that make up the biogroup and each of the queried biosets. The most significant biogroup receives a score of 100. All other biogroup scores are normalized to the top-ranked biogroup.¹⁸

Validation of Transcriptomic Data

Quantitative gene expression levels of relevant targets were confirmed using the QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific, Waltham, Massachusetts). Fold increases or decreases in gene expression were determined by quantitation of cDNA from tumor samples relative to control samples using the 2^−(ΔΔCt) method.

Immunohistochemistry for p44/42 MAPK (Erk1/2)

An immunohistochemical staining for p-MAPK was performed on sections of normal age-matched control lung tissue (n = 3) and AT-exposed ABCs (n = 7) from B6C3F1/N mice using the standard avidin-biotin-peroxidase technique. Formalin-fixed, paraffin-embedded tissue sections were deparaffinized in xylene and rehydrated through graded ethanol solutions. Heat-induced epitope retrieval was then performed using a citrate buffer (pH 6.0; Biocare Medical, Concord, California) in the Decloaker pressure chamber for 15 minutes at 110°C. Sections were then immersed in 3% H₂O₂ for 15 minutes to block endogenous peroxidase. Nonspecific sites were blocked by incubating slides for 20 minutes with 10% normal donkey serum (Jackson ImmunoResearch, West Grove, Pennsylvania), and endogenous biotin and avidin binding sites were blocked using an avidin-biotin blocking kit (Vector Laboratories, Burlingame, California). The sections were then incubated with rabbit polyclonal p-MAPK antibody (Cell Signaling Technology, Danvers, Massachusetts) at a 1:1000 dilution for 60 minutes at room temperature. For negative control tissue sections, normal rabbit IgG (MilliporeSigma, Burlington, Massachusetts) diluted to match the protein concentration of the p-MAPK antibody was utilized. Secondary incubation was done using a biotinylated donkey antirabbit IgG antibody (Jackson ImmunoResearch) at a dilution of 1:500 for 30 minutes at room temperature. Labeling incubation was done with the Vectastain RTU Kit Label (Vector Laboratories) for 30 minutes at room temperature. The antigen-antibody complex was visualized by staining with 3-diaminobenzidine chromogen (Agilent Dako, Santa Clara, California) for 6 minutes and counterstained with modified Harris hematoxylin.

Results

Kras and Egfr Mutations in AT-Exposed Mice and Rats

The incidences of Kras and Egfr mutations in alveolar/bronchiolar tumors in AT-exposed rats were 4% (1/26) and 50% (13/26), respectively (Table 3). The single Kras mutation was a non-hot-spot mutation on codon 16 with no established functional significance. The Egfr mutations were localized within exons 18 to 21 (G to A or C to T transitions). The incidence of Egfr mutations in alveolar/bronchiolar tumors from AT-exposed male rats was 62% compared to 39% in AT-exposed female rats.

Table 3.

Summary of Kras and Egfr mutations in nontumor lung tissue and alveolar/bronchiolar adenomas and carcinomas from Wistar Han rats after 2-year inhalation exposure to antimony trioxide.

Tissue:	Antimony trioxide	Mutation frequency^a		Kras codons	Egfr exons
Tissue:	Concentration	Kras	Egfr	16	18	19	20	21
Nontumor lung	0 mg/m³	0/11 (0)	0/11(0)	0	0	0	0	0
Lung tumors	0 mg/m³	0/4 (0)	0/4 (0)	0	0	0	0	0
	3 mg/m³	0/5 (0)	3^b/5 (60)	0	2	2	0	1
	10 mg/m³	1/11 (9)	6/11 (55)	1	2	2	1	1
	30 mg/m³	0/10 (0)	4/10 (40)	0	0	1	2	1
All exposed groups combined		1/26 (4)	13^b/26 (50)	1	4	5	3	3
Without pulmonary overload		0/5 (0)	3^b/5 (60)	0	2	2	0	1
With pulmonary overload		1/21 (5)	10/21 (48)	1	2	3	3	2

Male and female Wistar Han rats were exposed to 0, 3, 10, or 30 mg/m³ antimony trioxide daily by inhalation for 2 years. Silent mutations are not included. There were no significant differences from the chamber control group by the one-sided Fisher’s exact test and no significant exposure-concentration trend by the Cochran-Armitage trend test.

Number of tissues with mutation/number of tissues assayed (percentage with mutation).

Same animal with double mutations (M246 with Egfr 19 and 21; M257 with Egfr 18 and 19).

The incidence of Kras and Egfr mutations in alveolar/bronchiolar tumors in AT-exposed mice was 43% (34/80) and 46% (37/80), respectively (Table 4). A majority of the Kras mutations were located in codon 12 (G to A transitions). The Egfr mutations were mainly localized within exons 18 to 20 (G to A or C to T transitions). In alveolar/bronchiolar tumors from AT-exposed male mice, the incidence of Kras and Egfr mutations was 38% and 47%, respectively, compared to 49% and 46% in AT-exposed female mice.

Table 4.

Summary of Kras and Egfr mutations in nontumor lung tissue and alveolar/bronchiolar carcinomas from B6C3F1/N mice after 2-year inhalation exposure to antimony trioxide.

Tissue	Antimony trioxide	Mutation frequency^a		Kras codons					Egfr exons
Tissue	Concentration	Kras	Egfr	12	13	14	60	61	18	19	20	21
Nontumor lung	0 mg/m³	0/10 (0)	0/10 (0)	0	0	0	0	0	0	0	0	0
Lung tumors	0 mg/m³	3/9 (33)	0/9 (0)^#	3	0	0	0	0	0	0	0	0
	3 mg/m³	9/28 (32)	11/28 (39)*	4	0	0	1	4	5	0	4	2
	10 mg/m³	15/26 (58)	11^b/26 (42)*	11	0	0	0	4	2	1	8	1
	30 mg/m³	10/26 (38)	15^b/26 (58)**	8	1	1	0	0	4	7	4	1
All exposed groups combined		34/80 (43)	37^b/80 (46)**	23	1	1	1	8	11	8	16	4
Without pulmonary overload		9/28 (32)	11/28 (39)*	4	0	0	0	4	5	0	4	2
With pulmonary overload		25/52 (48)	26^b/52 (50)**	19	1	1	0	4	6	8	12	2

Male and female B6C3F1/N mice were exposed to 0, 3, 10, or 30 mg/m³ antimony trioxide daily by inhalation for 2 years. Silent mutations are not included.

Number of tissues with mutation/number of tissues assayed (percentage with mutation).

Same animal with double mutations (M449 with Egfr 20 and 21; M623 with Egfr 19 and 21).

Significantly different (P ≤ .05) from the chamber control group by the one-sided Fisher’s exact test.

P ≤ .005.

Significant exposure-concentration trend (P ≤ .01) by the Cochran-Armitage trend test.

Microarray Analysis

Principal component analysis was performed on all samples and all probes to characterize the variability present in the data. There was variability within the groups (control, spontaneous, AT-treated). The control group separated from the tumors, but the spontaneous and AT-treated tumors clustered together (Figure 2). Three samples (1 control [#5] and 2 spontaneous ABCs [#5 and #6]) that labeled poorly during microarray hybridization compared to the other samples were treated as outliers and removed from the analysis. Hierarchical cluster analysis (HCA) comparing standardized global gene expression profiles of all groups also showed clustering based on treatment, with some overlap between the spontaneous and AT-treated groups (Figure 3).

Figure 2.

Principal component analysis (PCA) showing the distribution of various sample types (control, spontaneous, AT-treated). Three samples (1 control and 2 spontaneous) that labeled poorly compared to the other samples were treated as outliers and removed from the analysis. ABC, alveolar/bronchiolar carcinomas (ABCs); AT, antimony trioxide; CNTL, control; PC, principal components; PCA, principal component analysis; SPNT, spontaneous ABCs.

Figure 3.

Hierarchical cluster analysis (HCA) comparing standardized global gene expression profiles (all 45,101 probe sets on the array) of normal lungs from control group, alveolar/bronchiolar carcinomas from the spontaneous group, and alveolar/bronchiolar carcinomas (ABCs) from the AT-treated group, in B6C3F1/N mice. AT, Antimony trioxide induced ABCs; SPNT, Spontaneous ABCs; CTL, Controls.

A total of 9,941 transcripts were differentially expressed in spontaneous tumors compared to control samples while 12,441 transcripts were altered in AT-exposed tumors versus control (FDR < 5%). Of this number, 7,716 transcripts were altered in both spontaneous tumors and AT-exposed tumors compared to control. In addition, 2,225 transcripts were unique to spontaneous tumors while 4,725 transcripts were unique to AT-exposed tumors (Figure 4).

Figure 4.

Venn diagram showing transcripts differentially expressed in spontaneous alveolar/bronchiolar carcinomas (ABCs) compared to control samples and antimony trioxide–induced ABCs compared to control samples and the corresponding significant overlap of differential expressed transcripts between the two tumor types.

Validation of Microarray Analysis (Real-Time PCR)

A subset of genes identified by microarray as dysregulated in AT-treated lung tumors were examined by TaqMan qPCR. In general, the expression of the genes examined by real-time PCR correlated well with results from microarray analysis.

Significantly Overrepresented Canonical Pathways in Mice AT-Exposed ABCs

Canonical pathways significantly altered in mouse AT-ABCs compared to controls (but not in spontaneous ABCs vs controls) as determined by Ingenuity pathway analyses of transcripts unique to AT-exposed ABCs are shown in Table 5. The ratio of number of genes within the differentially expressed gene list compared to the cataloged genes in that particular canonical pathway within the IPA database are shown, as well as −log(P value) as determined by Fisher’s exact test. Some relevant cancer pathways include cell cycle control of chromosomal replication (−log(P value) = 3.52), Ephrin receptor signaling (2.86), phospholipase C (PLC) signaling (2.79), signaling by Rho-family GTPases (2.58), paxillin signaling (2.27), and integrin-linked kinase (ILK) signaling (2.06).

Table 5.

Most significantly altered canonical molecular pathways in alveolar/bronchiolar carcinomas (ABCs) due to chronic antimony trioxide exposure in B6C3F1/N mice as determined by Ingenuity Pathway Analysis (IPA).

Canonical pathways	−log(P value)	Ratio^a
Cell cycle control of chromosomal replication	3.52	26/37 (0.703)
Ovarian cancer signaling	3.13	77/141 (0.546)
Pancreatic adenocarcinoma signaling	2.97	64/115 (0.557)
Ephrin receptor signaling	2.86	90/171 (0.526)
Phospholipase C signaling	2.79	110/215 (0.512)
Signaling by Rho-family GTPases	2.58	118/235 (0.502)
Mouse embryonic stem cell pluripotency	2.51	57/104 (0.548)
Paxillin signaling	2.27	57/106 (0.538)
Mismatch repair in eukaryotes	2.20	12/16 (0.750)
Purine nucleotides de novo biosynthesis II	2.14	9/11 (0.818)
Adipogenesis pathway	2.08	65/125 (0.520)
ILK (integrin-linked kinase) signaling	2.06	91/182 (0.500)
Glioma signaling	2.05	57/108 (0.528)

Statistical significance is set at P < .01 by Fisher’s exact test.

Ratio indicates the number of genes from that particular canonical pathway present in the differentially expressed gene data set compared to the total number of genes in that particular canonical pathway curated in the IPA database.

Significantly Upregulated and Downregulated Genes in AT-ABCs

Table 6 shows the top significantly (P < .001) upregulated and downregulated genes in AT-ABCs compared to controls. These genes were not significantly differentially expressed in spontaneous ABCs compared to controls. Top significantly upregulated genes include Dlk1 (fold change = +209.86), Tmem27 (+203.92), Rian (+66.39), Ctse (+44.96), Npy2r (+36.51), and Mycn (+12.83). Top significantly downregulated genes include Cyp2b6 (−112.085), Cbr2 (−23.29), Scgb1a1 (−21.67), Kcnk2 (−16.60), and Ripply3 (−13.77). Also, Ereg (P = .0019, fold change = +13.23) and Areg (P = .0053, fold change = +5.02), although not statistically significant at P < .001, were differentially expressed only in AT-exposed ABCs vs controls.

Table 6.

Top significantly upregulated and downregulated genes in alveolar/bronchiolar carcinomas (ABCs) from B6C3F1/N mice exposed to antimony trioxide for 2 years as determined by Ingenuity Pathway Analysis (IPA).

Gene	Description	Fold change	Relevance to cancer
Top twenty upregulated genes
Dlk1	Delta-like noncanonical Notch ligand 1	209.864	Highly expressed in NSCLC¹⁹; overexpression promotes lung cancer cell invasion²⁰
Tmem27	Transmembrane protein 27	203.919	Role unclear in cancer
Rian (Meg8)	RNA imprinted and accumulated in nucleus	66.389	RNA gene associated with EMT of lung and pancreatic cancer cells²¹
Ctse	Cathepsin E	44.964	Expression associated with pancreatic cancer²²; upregulated and hypomethylated in lung adenocarcinoma²³
Npy2r	Neuropeptide Y receptor Y2	36.513	Possible epigenetic marker in oral cancer²⁴
Serpinb5	Serpin family B member 5	33.006	Tumor-suppressing function in lung and breast cancer²⁵
Moxd1	Monooxygenase dopamine beta-hydroxylase (DBH) like 1	12.845	Role unclear in cancer
Mycn	v-myc Avian myelocytomatosis viral oncogene neuroblastoma-derived homolog	12.829	Documented oncogene; overexpressed in NSCLC²⁶
Fras1	Fraser extracellular matrix complex subunit 1	11.991	Migration and invasiveness of A549 cells reduced by Fras1 knockdown²⁷
Ptprn2	Protein tyrosine phosphatase, receptor type N2	9.463	Possible biomarker in lung cancer²⁸
Smarca1	SWItch/Sucrose Non-Fermentable (SWI/SNF)-related, matrix-associated, actin-dependent regulator of chromatin, subfamily a, member 1	8.232	Associated with proliferation and migration in colorectal cancer²⁹
Depdc7	Dishevelled, Egl-10 and Pleckstrin (DEP) domain containing 7	6.834	Found to inhibit cell proliferation and cell growth in hepatoma cells³⁰
Myom2	Myomesin 2	6.642	Role unclear in cancer
Cldn4	claudin 4	6.590	Overexpressed in ovarian cancer³¹ and pancreatic cancer³²
Cpne4	Copine 4	6.504	Role unclear in cancer
Ddc	Dopa decarboxylase	6.219	Role unclear in cancer
Scara3	Scavenger receptor class A member 3	6.091	Highly expressed in ovarian cancer³³
Mlana	Melan-A	5.670	Role unclear in cancer
Mcpt2	Mast cell protease 2	5.631	Role unclear in cancer
Hhex	Hematopoietically expressed homeobox	5.507	Role unclear in cancer
Top twenty downregulated genes
Cyp2b6	Cytochrome P450 family 2 subfamily B member 6	−112.085	Expressed in human lung;³⁴ may activate NNK³⁵
Cbr2	Carbonyl reductase 2	−23.290	Role unclear in cancer
Scgb1a1 (CC10)	Secretoglobin family 1A member 1	−21.666	May play a protective role in lung tumorigenesis induced by Nicotine-derived Nitrosamine Ketone (NNK), a potent carcinogen³⁶
Kcnk2	Potassium two-pore-domain channel subfamily K member 2	−16.596	Underexpressed in hepatocellular carcinoma cells³⁷
Ripply3	Ripply transcriptional repressor 3	−13.770	Role unclear in cancer
Sdpr	Serum deprivation response	−13.055	Suppresses metastasis in breast cancer³⁸; underexpressed in NSCLC³⁹
Cftr	Cystic fibrosis transmembrane conductance regulator	−12.483	Downregulated in NSCLC⁴⁰; DNA methylation may play role in downregulation of Cftr gene expression⁴¹
Cyp2e1	Cytochrome P450 family 2 subfamily E member 1	−12.311	Wild-type allele of Cyp2e1 associated with better survival in NSCLC patients⁴²
Cfap54	Cilia- and flagella-associated protein 54	−12.166	Role unclear in cancer
Hdgfrp3	Hepatoma-derived growth factor–related protein 3	−10.179	Upregulated in hepatocellular carcinoma⁴³
Pigr	Polymeric immunoglobulin receptor	−9.818	Downregulation associated with early lung tumorigenesis, overexpression of Pigr inhibits lung cancer cell proliferation⁴⁴
Ntn4	Netrin 4	−8.661	Role in suppressing liver metastasis of colorectal carcinoma⁴⁵
Cfap45 (CCDC19)	Cilia- and flagella-associated protein 45	−6.774	Potential tumor suppressor, downregulation associated with poor NSCLC prognosis⁴⁶
Mme	Membrane metalloendopeptidase	−6.116	High expression related to poor survival in NSCLC patients⁴⁷; underexpressed in lung adenocarcinoma⁴⁸
Irx2	Iroquois homeobox 2	−5.979	Role in suppressing chemokine expression and cell motility in breast cancer⁴⁹
Fam65b	Family with sequence similarity 65 member B	−5.965	Overexpression can block proliferation of transformed and primary T cells⁵⁰
Dnaja4	DnaJ heat shock protein family (Hsp40) member A4	−5.462	Role unclear in cancer
Il2rj	Interleukin 2 receptor subunit gamma	−5.208	Role unclear in cancer
Ifit3	Interferon-induced protein with tetratricopeptide repeats 3	−5.187	Inhibits growth of lung cancer cells by suppressing STAT3⁵¹
Rragd	Ras-related GTP binding D	−5.067	Found to be downregulated in melanoma metastasis samples⁵²

Statistical significance is set at P < .001 by Fisher’s exact test.

Abbreviations: EMT, epithelial-mesenchymal transition; NSCLC, non-small-cell lung cancer.

Genes and Pathways in Mouse AT-Exposed ABCs Similar to Those in Human NSCLC (Illumina BaseSpace Correlation Engine)

Using Illumina BaseSpace Correlation Engine, the differentially expressed transcripts data set of mouse AT-ABCs was compared to the most similar human NSCLC data set (from the Illumina BaseSpace Correlation Engine curated data sets as well as the National Center of Biotechnology Information’s GEO database). The human NSCLC data set (GSE27262) that was most concordant with the mouse AT-exposed ABC tumor data set was from stage-I lung adenocarcinoma from patients compared to adjacent nontumor tissue using Affymetrix Human Genome U133 Plus 2.0 Array.

Comparisons of the 8,067 differentially expressed transcripts in AT-ABCs in mice to the 12,667 differentially expressed human NSCLC genes resulted in an overlap of 3,953 genes (P = 1.0E-48), with 1,735 upregulated genes (P = 3.9E-23) and 1,181 downregulated genes (P = 1.6E-78) (Figure 5). Top concordant upregulated genes in AT-ABCs in mice and human NSCLC include Dlk1, Tmem27, Bpifa1 and Ros1, and the top concordant downregulated genes include Ogn, Gpm6a, Mfap4, and Scn7a (Table 7). Top concordant canonical pathways in mouse AT-ABCs in mice and human NSCLC include Nop56p-associated pre-rRNA complex, 60S ribosomal subunit (cytoplasmic), 55S ribosome (mitochondrial), 40S ribosomal subunit (cytoplasmic), and oxidative phosphorylation (Table 8).

Figure 5.

Illumina BaseSpace Correlation Engine was used to compare the differentially expressed mouse antimony trioxide (AT)-exposed ABCs tumor data set to the most similar human NSCLC data set (from the Illumina BaseSpace curated data sets as well as National Center for Biotechnology Information’s Gene Expression Omnibus (GEO) database). The human NSCLC data set (GSE27262) that closely matches the mouse AT-ABC tumor data set was from stage-I lung adenocarcinoma from patients compared to adjacent nontumor tissue using Affymetrix Human Genome U133 Plus 2.0 Array. ABC, alveolar/bronchiolar carcinoma; NSCLC, non-small-cell lung cancer.

Table 7.

The top concordant upregulated and downregulated genes in ABCs from B6C3F1/N mice exposed to antimony trioxide for 2 years and human non-small-cell lung cancer as determined by Illumina BaseSpace Correlation Engine.

Gene	Description	Mouse AT ABCs	Mouse spontaneous ABCs	Human NSCLC	Known relevance to cancer
Top twenty upregulated genes
Dlk1	Delta-like 1 homolog (Drosophila)	209.864	—	1.67	Overexpression promotes lung cancer cell invasion²⁰
Tmem27	Transmembrane protein 27	203.919	—	2.18	Role in cancer unclear
Bpifa1 (Lunx)	Bactericidal or Permeability-Increasing protein (BPI) fold-containing family A, member 1	86.890	186.552	1.91	Expression is significantly higher in NSCLC vs noncancerous tissue⁵³
Ros1	Ros1 proto-oncogene	34.188	32.183	2.63	Gene rearrangements occurs in approximately 1% to 2% of NSCLC⁵⁴
Trpv6	Transient receptor potential cation channel, subfamily V, member 6	29.483	20.227	1.31	Trpv6 and Trpv5 expression are reduced in tumor issues of NSCLC patients⁵⁵
Bcl2l15	BCLl2-like 15	29.296	—	5.24	Role in cancer unclear
Sertad4	SERTA domain containing 4	29.229	35.867	1.63	Role in cancer unclear
Sox9	SRY-box containing gene 9	25.855	20.302	4.11	Promotes epithelial-mesenchymal transition in NSCLC⁵⁶; possible biomarker for NSCLC⁵⁷
Cldn2	Claudin 2	23.537	23.379	4.00	Promote promatrix metalloproteinases, involved in epithelial-mesenchymal transition⁵⁸
Pkhd1	Polycystic kidney and hepatic disease 1	18.251	6.152	1.73	Role in cancer unclear
Pthlh	Parathyroid hormone-like peptide	18.096	20.908	1.22	Mediates lung metastasis of KRAS-mutated colorectal cancer cells⁵⁹
S100a14	S100 calcium-binding protein A14	17.717	7.285	1.86	Expressed in a subset of human lung adenocarcinomas⁶⁰
Arg2	Arginase type II	16.942	22.408	1.52	Significantly expressed in human small-cell lung cancers
Stk39	Serine/threonine kinase 39	15.619	12.633	3.08	Function as tumor oncogene in NSCLC⁶¹
Bex2	Brain expressed X-linked 2	14.330	—	1.74	Promotes proliferation of glioblastoma cells⁶² and colorectal cancer cells⁶³
Nipal1	NIPA-like domain containing 1	12.989	8.263	1.46	Highly expressed in oral squamous cell carcinoma, possible oncogene⁶⁴
Zfp618	Zinc finger protein 618	12.774	—	1.27	Role in cancer unclear
Celf4	CUGBP, Elav-like family member 4	12.574	6.927	1.27	Role in cancer unclear
Slc15a1	Solute carrier family 15 (oligopeptide transporter), member 1	12.192	11.604	1.46	Role in cancer unclear
Slc15a2	Solute carrier family 15 (H+/peptide transporter), member 2	11.443	9.180	1.42	Role in cancer unclear
Top twenty downregulated genes
Ogn	Osteoglycin	−170.083	−126.005	−8.21	Tumor-suppressing role in breast⁶⁵ and colorectal cancer⁶⁶
Gpm6a	Glycoprotein m6a	−145.801	−170.789	−28.3	Role in cancer unclear
Mfap4	Microfibrillar-associated protein 4	−89.389	−88.929	−6.37	Involved in cell adhesion, potential tumor suppressor⁶⁷
Scn7a	Sodium channel, voltage-gated, type VII, alpha	−83.597	−68.427	−4.90	Significantly associated with survival in lung cancer patients⁶⁸
1110017D15Rik	RIKEN cDNA 1110017D15 gene	−61.358	—	−2.50	Role in cancer unclear
4833427G06Rik	RIKEN cDNA 4833427G06 gene	−60.671	—	−2.02	Role in cancer unclear
Pcolce2	Procollagen C-endopeptidase enhancer 2	−56.017	−28.611	−7.72	Role in cancer unclear
Tmem100	Transmembrane protein 100	−53.002	−57.658	−24.9	Potential tumor suppressor in human NSCLC⁶⁹
Cyp4b1	Cytochrome P450, family 4, subfamily b, polypeptide 1	−51.096	−28.929	−4.80	Downregulated in human NSCLC and other lung tumors⁷⁰
Fmo3	Flavin-containing monooxygenase 3	−49.629	−43.592	−2.66	Role in cancer unclear
Ntrk2	Neurotrophic tyrosine kinase, receptor, type 2	−37.760	−39.431	−2.52	Significantly downregulated in lung adenocarcinoma⁷¹
AU021034	Expressed sequence AU021034	−37.132	—	−2.71	Role in cancer unclear
Efcab1	EF hand calcium-binding domain 1	−35.704	−30.059	−1.83	Role in cancer unclear
Tubb1	Tubulin, beta 1 class VI	−34.052	−34.886	−3.00	Role in cancer unclear
Dpt	Dermatopontin	−33.766	−18.368	−4.16	Overexpression of Dpt suppresses hepatocellular carcinoma cell migration⁷²
Inmt	Indolethylamine N-methyltransferase	−31.569	−27.623	−3.55	Role in cancer unclear
Figf	c-fos-induced growth factor	−30.376	—	−12.7	Downregulated in lung squamous cell carcinoma⁷³
Ppbp	Pro-platelet basic protein	−28.726	−19.302	−10.3	Role in cancer unclear
Hpgd	Hydroxyprostaglandin dehydrogenase 15 (NAD)	−28.194	−25.948	−2.38	Tumor suppressor function, overexpression can induce apoptosis⁷⁴
Spock2	sparc/osteonectin, cwcv, and kazal-like domains proteoglycan 2	−26.545	−12.428	−7.20	Upregulation of Spock2 inhibits migration and invasion of prostate cancer cells⁷⁵

Corresponding genes, if applicable, in spontaneously arising ABCs from B6C3F1/N mice are shown for comparison.

Abbreviations: ABCs, alveolar/bronchiolar carcinomas; NSCLC, non-small-cell lung cancer, AT, Antimony trioxide.

Table 8.

Illumina BaseSpace Correlation Engine generated top canonical pathways common to alveolar/bronchiolar carcinomas (ABCs) from B6C3F1/N mice exposed to antimony trioxide (AT) for 2 years and human non-small-cell lung cancer as predicted by the Broad Institute’s Molecular Signatures Database (MSigDB) using the respective transcriptomic data.

Top 20 concordant canonical pathways in mouse AT-exposed ABCs and human NSCLC	Bioset score	Mouse, −log(P value)	Common genes	Human, −log(P value)	Common genes
Nop56p-associated pre-rRNA complex	93	40.24	76	0.04	18
60S ribosomal subunit, cytoplasmic	68	29.60	42	0.04	8
55S ribosome, mitochondrial	56	24.15	52	17.64	49
40S ribosomal subunit, cytoplasmic	49	21.32	29	0.03	3
Oxidative phosphorylation	48	20.85	60	8.59	52
Huntington’s disease	43	18.49	72	4.47	70
Focal adhesion	40	17.38	65	26.64	97
Vascular smooth muscle contraction	38	16.40	45	20.29	61
Parkinson’s disease	35	15.36	52	5.46	47
39S ribosomal subunit, mitochondrial	35	15.12	32	9.89	29
Validated targets of C-MYC (Mylocytomatosis) transcriptional activation	35	15.07	41	12.08	41
Cytokine-cytokine receptor interaction	34	14.96	63	19.66	91
Hematopoietic cell lineage	34	14.74	33	12.43	42
Pyrimidine metabolism	29	12.74	42	13.62	49
Alzheimer’s disease	27	11.89	60	3.30	58
Chemokine signaling pathway	25	10.74	49	19.41	83
Purine metabolism	24	10.54	53	11.00	64
Lysophosphatidic acid (LPA) receptor mediated events	24	10.48	27	8.92	31
DNA replication	23	10.10	21	11.30	25
28S ribosomal subunit, mitochondrial	23	9.77	20	8.74	20

Abbreviations: NSCLC, non-small-cell lung cancer.

Immunohistochemistry for p44/42 MAPK (Erk1/2)

In normal control lungs, nuclear expression of p44/42 MAPK was present in the endothelium, but the pulmonary epithelial cells did not exhibit immunoreactivity for pErk1/2 (Figure 6A and B). In ABCs with Kras mutations, pErk1/2 expression was predominantly nuclear with some cytoplasmic expression. Expression was mainly in the neoplastic cells at the periphery of the neoplasm and in the neoplastic cells that infiltrated the surrounding parenchyma (Figure 6C and D).

Figure 6.

Expression of p44/42 MAPK in B6C3F1/N mouse lung. In normal mouse lungs, p44/42 MAPK expression is predominantly in the endothelium (A and B). However, the neoplastic alveolar/bronchiolar cells stain positive for p44/42 MAPK in the mouse lung tumors. The distribution pattern of p44/42 MAPK expression in neoplastic cells was predominantly nuclear and mainly in neoplastic cells at the periphery of the neoplasm and those infiltrating the surrounding parenchyma (C and D). MAPK, mitogen-activated protein kinase. To access the whole-slide images provided with this figure, scan the QR code on cover page 3 or visit slide 1 , slide 2 , slide 3 .

Discussion

Alveolar/bronchiolar tumors resulting from exposure to AT in both mice and rats harbored mutations in the Egfr gene. However, Kras mutations (mainly G to A transitions in codon 12) were observed only in alveolar/bronchiolar tumors in B6C3F1/N mice, but not in rats, exposed to AT. These Kras mutations, however, were also common in spontaneous ABCs,¹³ suggesting AT exposure may have enhanced proliferation of neoplastic clones harboring the spontaneous Kras mutations in mice.

KRAS mutations in NSCLC are seen more frequently in smokers, while EGFR mutations in lung cancers are most frequently observed in never-smokers, Asians, and women.⁵ In general, mutations within EGFR and KRAS are considered to be mutually exclusive events in human lung and colon cancers.^76
-79 Interestingly in this study, 32% (15/47) of the ABCs from AT-exposed mice harbored both Egfr and Kras mutations. In AT-exposed rats, only one ABC harbored mutations in both Egfr and Kras (a non–hot spot mutation on codon 16 with unknown functional significance). The occurrence of concurrent mutations was also noted in ABCs in rats and mice resulting from chronic exposure to cobalt metal dust.⁹ These data suggest that chronic chemical exposure may lead to multiple simultaneous mutational events leading to heterogenous clonal evolution.

Another interesting feature of this study is the occurrence of pulmonary overload in rats and mice at 10 and 30 mg/m³ but not at 3 mg/m³ (Tables 3 and 4). The concept of pulmonary overload implies that the exposure to high levels of poorly soluble particles exceeding the clearance capacity of the lung leads to a carcinogenic response unrelated to the intrinsic toxicity of the exposure, that is, lung overload due to inert particles such as talc and carbon black cause pulmonary inflammation, fibrosis, epithelial hyperplasia, and eventually lung tumors.⁸⁰ In this study, the incidences of lung tumors as well as the incidences or types of Kras or Egfr mutations in lung tumors from mice and rats were not significantly different in exposures that did and did not approach the pulmonary overload threshold. This suggests that the mutation spectra of these neoplasms were not affected by pulmonary overload and are probably related to the intrinsic toxicity of AT exposure.

Immunochemical demonstration of p44/42 MAPK expression in FFPE sections of ABCs arising from AT-exposure in mice that harbored Kras and/or Egfr mutations showed upregulation of the downstream MAPK signaling. The upregulation seems to be predominantly in the neoplastic cells at the periphery of the neoplasm or in the infiltrating neoplastic cells in the surrounding parenchyma. Moreover, the staining pattern was similar in ABC sections with a Kras mutation or both Kras and Egfr mutations. The role of ERK 1/2 or MAPK pathway activity in epithelial-mesenchymal transition (EMT) was previously demonstrated by several studies both in transformed and nontransformed epithelial cells.^81
-84 The expression pattern of phosphorylated Erk1/2 in mouse ABCs harboring mutations in Kras or both Kras and Egfr is predominantly restricted to the periphery of the neoplastic mass, supporting possible involvement in EMT and tumor cell migration.

Differential transcriptomic analysis of B6C3F1/N ABCs in this study showed several significantly dysregulated canonical pathways relevant to carcinogenesis. Alveolar/bronchiolar adenomas/carcinomas that arise spontaneously or from chronic chemical exposures are morphologically indistinguishable.¹³ In this study, the PCA and HCA plots showed overlap between the ABCs arising spontaneously or due to chronic AT exposure, suggesting shared molecular events leading to tumorigenesis and tumor progression between these etiologies. However, there were some significantly altered molecular changes and canonical pathways unique to AT-exposed ABCs.

Several canonical pathways were significantly altered in AT-exposed ABCs but not in spontaneous ABCs (Table 5), including cancer-relevant pathways such as ephrin receptor signaling (−log(P value) = 2.86), PLC signaling (2.79), signaling by Rho-family GTPases (2.58), paxillin signaling (2.27), and ILK signaling (2.06). Eph receptors are transmembrane proteins (TMEMs) involved in organismal growth and development, including processes such as axon guidance, neural development, and vascular development.⁸⁵ They are activated by a family of cell surface–associated ligands called ephrins. Aberrant expression of ephrin receptor signaling is found to contribute to various cancers.⁸⁵ Several Eph receptor genes were altered in ABCs of AT-exposed mice. EphA2 is overexpressed in NSCLC,⁸⁶ and its expression also correlates with brain metastasis in NSCLC patients.⁸⁷ The silencing of EphA7 is shown to inhibit proliferation and metastasis of A549 human lung cancer cells.⁸⁸ Eph receptors are also negative regulators of MAPK signaling.⁸⁹ EphrinA1 can suppress MAPK signaling and the proliferation of prostatic cancer cells.⁹⁰ It is possible that aberrant ephrin receptor signaling plays a role in alteration of MAPK signaling in AT-exposed ABCs. Phospholipase C signaling plays a role in intracellular and second messenger signaling. It consists of six classes of PLC. One of these classes, PLC-β, is found to mediate the proliferation of NSCLC and SCLC cells.⁹¹ In the canonical pathway signaling by Rho-family GTPases, three subfamilies of small GTP-binding proteins, RHO, RAC, and CDC42, play the primary roles in mammals in the diverse biological processes regulated by the pathway, including cytoskeleton formation and transcription regulation.⁹² RAC and CDC42 can also activate the PI3K/AKT signaling pathway, which is dysregulated in many cancers. Rho GTPases are aberrantly expressed in various human cancers, including breast, lung, and colon cancers.⁹² RhoA is found to be overexpressed in lung cancers,⁹³ and RhoE is overexpressed in NSCLC.⁹⁴ Transformation of cells by oncogenic Ras also requires crosstalk between Ras and Rho-family GTPases signaling.⁹⁵ Rho signaling can block the growth inhibitory effects of ERK-MAPK signaling. In turn, ERK-MAPK signaling can block RhoA signaling to the cytoskeleton, effectively promoting motility and proliferation of transformed cells.⁹⁵ Paxillin is a focal adhesion-associated protein that regulates cell spreading and motility.⁹⁶ In NSCLC cell lines, paxillin overexpression was found to inhibit cell motility.⁹⁷ Integrin-linked kinase was shown in vitro to promote NSCLC development and metastasis through increased cell migration and invasion by inducing EMT.⁹⁸ In patients with EGFR-mutation-positive NSCLC, high levels of mRNA expression for ILK point toward shorter overall survival.⁹⁹

Several genes were significantly differentially expressed only in AT-exposed ABCs (Table 6). Dlk1, delta like noncanonical Notch ligand 1, is markedly upregulated (+209.86 fold change) in AT-ABCs. DLK1 is highly expressed in NSCLC,¹⁹ and its overexpression can promote lung cancer cell invasion.²⁰ Transmembrane proteins are a group of 310 proteins that are components of cellular membranes. TMEMs are noted to function as both tumor suppressors and oncogenes in various cancers.¹⁰⁰ Tmem27 is markedly upregulated (+203.92) in AT-exposed ABCs. The role of TMEMs in AT-induced carcinogenesis is unclear. Rian (+66.39), RNA imprinted and accumulated in nucleus (also known as Meg8), encodes for a long noncoding RNA expressed in the nucleus. It is associated with EMT in lung and pancreatic cancer cells²¹ and the tumor progression of NSCLC.¹⁰¹ Ctse (+44.96), cathepsin E, is a member of the peptidase A1 proteases and is found in various cell types, such as cells of the gastrointestinal and immune systems and also in cancer cells.²² Its expression is associated with pancreatic cancer,²² and it is hypomethylated and upregulated in lung adenocarcinoma.²³ Mycn, v-myc avian myelocytomatosis viral oncogene neuroblastoma-derived homolog, is a well-documented oncogene and is upregulated in AT-exposed ABCs (+12.83). Mycn is overexpressed in NSCLC and was found to promote NSCLC cell proliferation.²⁶ Scgb1a1 (−21.67), secretoglobin family 1A member 1, is a secreted protein and member of the secretoglobin family. It is associated with anti-inflammation, and defects in Scgb1a1 are associated with asthma susceptibility. Mice with this gene knockout and exposed to the carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) showed markedly increased frequency of Kras mutations and alterations of MAPK/Erk1 signaling in lung tumors.³⁶

Commonly altered canonical pathways between mice AT-exposed ABCs and human NSCLC (as determined by Illumina BaseSpace in Table 8) include 55S ribosome, mitochondrial (-log(P value) = 24.15 in AT-ABCs and 17.64 in human NSCLC), oxidative phosphorylation (20.85 in AT-ABCs and 8.59 in human NSCLC), focal adhesion (17.38 in AT-ABCs and 26.64 in human NSCLC), and vascular smooth muscle contraction (16.40 in AT-ABCs and 20.29 in human NSCLC). Several differentially expressed genes that were statistically significant and highly altered (±fold changes) in mouse AT-ABCs were also altered in human NSCLCs (Table 7). Dlk1 and Tmem27 were already discussed previously in genes significantly differentially expressed only in AT-exposed ABCs as determined by IPA. Osteoglycin (Ogn) is part of the small leucine-rich proteoglycans in the extracellular matrix and is abundantly expressed in the mouse lung. Ogn is markedly downregulated both in AT-exposed ABCs (−170.083) and in human NSCLC (−8.21). Deficiency in OGN is related to defects in collagen fibril morphology.¹⁰² OGN plays a tumor suppressor role in human breast cancer where it inhibits cell proliferation and invasiveness by suppressing PI3K/Akt/mTOR signaling.⁶⁵ It was shown that decreased OGN expression allowed tumor cells to induce EGFR activity to sustain proliferative signaling in colorectal cancer.⁶⁶ Microfibrillar-associated protein 4 (Mfap4) is an extracellular matrix protein the binds collagen and contains a C terminal fibrinogen-like domain and an N-terminal integrin-binding motif. Mfap4 is involved in cell adhesion and intercellular interactions and is downregulated in both murine and human lung tumors. Mfap4 is downregulated in AT-exposed ABCs (−89.389) and human NSCLC (−6.37). The mRNA level of MFAP4 is significantly downregulated in cancer tissues compared with normal tissues, indicating that MFAP4 may potentially be a tumor suppressor gene.⁶⁷

There were over twice as many unique genes (4,725 transcripts) that were significantly differentially expressed in ABCs in mice exposed to AT compared with those (2,225 transcripts) from spontaneously arising ABCs. Several noted cancer genes such as Ereg (+13.2), Areg (+5.0), and Mycn (+12.9) related to MAPK signaling and lung cancer were significantly differentially altered in the AT-exposed ABCs but not in ABCs arising spontaneously. It has been shown that EREG, a ligand of EGFR, is required for the promotion of lung tumors.¹⁰³ It has also been shown that oncogenic mutations in the Kras and Egfr genes activate the MEK/ERK pathway and induce overexpression of Ereg. Consequently, Ereg overexpression plays an essential role in oncogenic KRAS-mediated tumorigenesis.¹⁰⁴ It is, therefore, possible that Ereg also plays a role in the tumorigenesis of ABCs of mice exposed to AT through oncogenic activation of the MEK/ERK pathway. Another ligand of EGFR, AREG, is secreted by NSCLCs. AREG has been shown to promote autonomous growth of tumor cells and to provide resistance to apoptosis.¹⁰⁵ Areg is significantly overexpressed in ABCs resulting from AT exposure but not differentially expressed in spontaneous ABCs compared to normal lung. Therefore, Areg, like Ereg, may also play a role in the progression of ABCs in AT-exposed mice. Mycn is overexpressed in AT-treated ABCs. It has been shown that MYCN is overexpressed in NSCLC in humans and may also play a role in promoting the proliferation of these cancer cells.²⁶ Myc/Mycn is a target molecule in the MAPK signaling pathway. Ingenuity Pathway Analysis identified Mycn as a predicted upstream regulator in AT-treated and spontaneous alveolar/bronchiolar tumors.

Recently, Riva et al¹⁰⁶ examined the whole genomes of a subset of mouse ABCs arising spontaneously or due to chronic exposure AT obtained from this study and demonstrated that the mutation burden and mutation signatures (single-base substitutions [SBS], doublet mutations, and small insertions and deletion) are comparable. The main difference being slightly increased SBS18 signature in ABCs arising from AT exposure, indicating increased oxidative stress. Differential transcriptomic analysis of these ABCs suggests shared molecular events leading to tumorigenesis and tumor progression in both groups of ABCs, with the exception of some significantly differentially expressed genes and altered canonical pathways unique to AT-exposed ABCs. Based on the data from the study by Riva et al¹⁰⁶ and this study, it appears that AT-exposure potentiates some of these pathways in spontaneous ABCs as observed by increased tumor incidences and tumor multiplicities in ABCs after chronic AT-exposure.

In summary, this study shows that mice and rats exposed to AT develop alveolar/bronchiolar tumors that harbor Egfr mutations (mice and rats) and Kras mutations (mice). Hot spot mutation data and transcriptomic analysis of these lung tumors show that alterations in the MAPK signaling are important in AT-induced pulmonary carcinogenesis. Alterations in MAPK signaling are also important in human lung cancers, demonstrating the translational relevance of these tumors in human health risk.

Footnotes

Acknowledgements

We would like to dedicate this manuscript to the memory of Dr. Gordon Flake, our trusted and knowledgeable colleague who passed away after completing this technical report (NTP TR 590). We appreciate the help provided by the National Toxicology Program Tissue Archives, the National Institute of Environmental Health Sciences (NIEHS) pathology core laboratories, as well as the NIEHS Microarray core laboratory. In addition, we express our gratitude to Drs. Jian-Liang-Li and Alex Merrick for reviewing the manuscript and providing feedback.

Abbreviations

ABA, alveolar/bronchiolar adenoma; ABC, alveolar/bronchiolar carcinoma; AT, antimony trioxide; FFPE, formalin-fixed paraffin-embedded; HCA, hierarchical cluster analysis; IPA, ingenuity pathway analysis; NTP, national toxicology program; PCA, principal component analysis; RMA, robust multiarray analysis; SPNT, spontaneous.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research has been supported by the DNTP/NIEHS intramural program (ES103319-06 [2021], ES102505-14 [2021], and ZIA ES103383-01).

ORCID iDs

Ramesh C. Kovi

Kyathanahalli S. Janardhan

Arun R. Pandiri

References

National Toxicology Program (NTP). Report on Carcinogens monograph on antimony trioxide. Research Triangle Park, NC: National Toxicology Program. RoC Monograph 13. 2018. https://doi.org/10.22427/ROC-MGRAPH-13

National Research Council. Toxicological Risks of Selected Flame-Retardant Chemicals. Washington, DC: National Academy Press; 2000.

Hazardous Substances Data Bank. Antimony trioxide database search. PubChem (nih.gov). Published 2013.

National Toxicology Program. NTP technical report on the toxicology and carcinogenesis studies of antimony trioxide (CASRN 1309-64-4) in Wistar Han [CRL:WI(HAN)] rats and B6C3F1/N mice (inhalation studies). Toxic Rep Ser. 2017;590:1-L3.

Boch

Kollmeier

Roth

, et al. The frequency of EGFR and KRAS mutations in non-small cell lung cancer (NSCLC): routine screening data for central Europe from a cohort study. BMJ Open. 2013;3(4):e002560. doi:10.1136/bmjopen-2013-002560.

Ellis

Clark

The importance of being K-Ras. Cell Signal. 2000;12(7):425-434.

Roberts

Der

CJ.

Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer. Oncogene. 2007;26(22):3291-3310. doi:10.1038/sj.onc.1210422.

Johnson

Janne

PA.

Epidermal growth factor receptor mutations in patients with non-small cell lung cancer. Cancer Res. 2005;65(17):7525-7529. doi:10.1158/0008-5472.CAN-05-1257.

Hong

Hoenerhoff

Ton

, et al. Kras, Egfr, and Tp53 mutations in B6C3F1/N mouse and F344/NTac rat alveolar/bronchiolar carcinomas resulting from chronic inhalation exposure to cobalt metal. Toxicol Pathol. 2015;43(6):872-882. doi:10.1177/0192623315581192.

10.

Ton

Kovi

Peddada

, et al. Cobalt-induced oxidative stress contributes to alveolar/bronchiolar carcinogenesis in B6C3F1/N mice. Arch Toxicol. 2021;95(10):3171-3190. doi:10.1007/s00204-021-03146-5.

11.

Rodenhuis

van

Wetering

Mooi

Evers

van Zandwijk

Bos

JL.

Mutational activation of the K-ras oncogene. A possible pathogenetic factor in adenocarcinoma of the lung. N Engl J Med. 1987;317(15):929-935. doi:10.1056/NEJM198710083171504.

12.

Midha

Dearden

McCormack

EGFR mutation incidence in non-small-cell lung cancer of adenocarcinoma histology: a systematic review and global map by ethnicity (mutMapII). Am J Cancer Res. 2015;5(9):2892-2911.

13.

Hong

Ton

Kim

, et al. Genetic alterations in K-ras and p53 cancer genes in lung neoplasms from B6C3F1 mice exposed to cumene. Toxicol Pathol. 2008;36(5):720-726. doi:10.1177/0192623308320280.

14.

Irizarry

Hobbs

Collin

, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003;4(2):249-264. doi:10.1093/biostatistics/4.2.249.

15.

Peddada

Harris

Zajd

Harvey

ORIOGEN: order restricted inference for ordered gene expression data. Bioinformatics. 2005;21(20):3933-3934. doi:10.1093/bioinformatics/bti637.

16.

Ward

Hierarchical grouping to optimize an objective function. J Am Stat Assoc. 1963;58:236-244.

17.

Subramanian

Tamayo

Mootha

, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545-15550. doi:10.1073/pnas.0506580102.

18.

Kupershmidt

Grewal

, et al. Ontology-based meta-analysis of global collections of high-throughput public data. PLoS ONE. 2010;5(9):e13066. doi:10.1371/journal.pone.0013066.

19.

Liu

Tan

, et al. [Study on the molecular mechanisms of dlk1 stimulated lung cancer cell proliferation]. Zhongguo Fei Ai Za Zhi. 2010;13(10):923-927. doi:10.3779/j.issn.1009-3419.2010.10.12.

20.

Tan

Zhang

, et al. DLK1 promotes lung cancer cell invasion through upregulation of MMP9 expression depending on Notch signaling. PLoS ONE. 2014;9(3):e91509. doi:10.1371/journal.pone.0091509.

21.

Terashima

Ishimura

Wanna-Udom

Suzuki

MEG8 long noncoding RNA contributes to epigenetic progression of the epithelial-mesenchymal transition of lung and pancreatic cancer cells. J Biol Chem. 2018;293(47):18016-18030. doi:10.1074/jbc.RA118.004006.

22.

Pontious

Kaul

Hong

, et al. Cathepsin E expression and activity: role in the detection and treatment of pancreatic cancer. Pancreatology. 2019;19(7):951-956. doi:10.1016/j.pan.2019.09.009.

23.

Yang

Wang

Liu

Exploring and comparing of the gene expression and methylation differences between lung adenocarcinoma and squamous cell carcinoma. J Cell Physiol. 2019;234(4):4454-4459. doi:10.1002/jcp.27240.

24.

Misawa

Imai

Mochizuki

, et al. Genes encoding neuropeptide receptors are epigenetic markers in patients with head and neck cancer: a site-specific analysis. Oncotarget. 2017;8(44):76318-76328. doi:10.18632/oncotarget.19356.

25.

Chou

Wen

Liang

, et al. Suppression of the invasion and migration of cancer cells by SERPINB family genes and their derived peptides. Oncol Rep. 2012;27(1):238-245. doi:10.3892/or.2011.1497.

26.

Liu

Wang

Liu

, et al. Overexpression of MYCN promotes proliferation of non-small cell lung cancer. Tumour Biol. 2016;37(9):12855-12866. doi:10.1007/s13277-016-5236-2.

27.

Zhan

Huang

Liang

, et al. FRAS1 knockdown reduces A549 cells migration and invasion through downregulation of FAK signaling. Int J Clin Exp Med. 2014;7(7):1692-1697.

28.

Wielscher

Vierlinger

Kegler

Ziesche

Gsur

Weinhäusel

Diagnostic performance of plasma DNA methylation profiles in lung cancer, pulmonary fibrosis and COPD. EBioMedicine. 2015;2(8):929-936. doi:10.1016/j.ebiom.2015.06.025.

29.

Liu

Han

Zhu

Sun

Zhu

LncRNA DLEU1 contributes to colorectal cancer progression via activation of KPNA3. Mol Cancer. 2018;17(1):118. doi:10.1186/s12943-018-0873-2.

30.

Liao

Wang

Lin

DEPDC7 inhibits cell proliferation, migration and invasion in hepatoma cells. Oncol Lett. 2017;14(6):7332-7338. doi:10.3892/ol.2017.7128.

31.

Choi

Kim

Kwon

, et al. Expression profile of tight junction protein claudin 3 and claudin 4 in ovarian serous adenocarcinoma with prognostic correlation. Histol Histopathol. 2007;22(11):1185-1195. doi:10.14670/HH-22.1185.

32.

Michl

Buchholz

Rolke

, et al. Claudin-4: a new target for pancreatic cancer treatment using clostridium perfringens enterotoxin. Gastroenterology. 2001;121(3):678-684. doi:10.1053/gast.2001.27124.

33.

Bock

Nymoen

Brenne

Kærn

Davidson

SCARA3 mRNA is overexpressed in ovarian carcinoma compared with breast carcinoma effusions. Hum Pathol. 2012;43(5):669-674. doi:10.1016/j.humpath.2011.06.003.

34.

Hukkanen

Pelkonen

Hakkola

Raunio

Expression and regulation of xenobiotic-metabolizing cytochrome P450 (CYP) enzymes in human lung. Crit Rev Toxicol. 2002;32(5):391-411. doi:10.1080/20024091064273.

35.

Dicke

Skrlin

Murphy

SE.

Nicotine and 4-(methylnitrosamino)-1-(3-pyridyl)-butanone metabolism by cytochrome P450 2B6. Drug Metab Dispos. 2005;33(12):1760-1764. doi:10.1124/dmd.105.006718.

36.

Yang

Zhang

Mukherjee

Linnoila

RI.

Increased susceptibility of mice lacking Clara cell 10-kDa protein to lung tumorigenesis by 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone, a potent carcinogen in cigarette smoke. J Biol Chem. 2004;279(28):29336-29340. doi:10.1074/jbc.C400162200.

37.

Xiong

Huang

, et al. KCNK levels are prognostic and diagnostic markers for hepatocellular carcinoma. Aging (Albany NY). 2019;11(19):8169-8182. doi:10.18632/aging.102311.

38.

Ozturk

Papageorgis

Wong

, et al. SDPR functions as a metastasis suppressor in breast cancer by promoting apoptosis. Proc Natl Acad Sci U S A. 2016;113(3):638-643. doi:10.1073/pnas.1514663113.

39.

Zhu

Wang

Liao

Long non-coding RNA SDPR-AS affects the development of non-small cell lung cancer by regulating SDPR through p38 MAPK/ERK signals. Artif Cells Nanomed Biotechnol. 2019;47(1):3172-3179. doi:10.1080/21691401.2019.1642904.

40.

Zhang

Jiang

, et al. The cystic fibrosis transmembrane conductance regulator as a biomarker in non-small cell lung cancer. Int J Oncol. 2015;46(5):2107-2115. doi:10.3892/ijo.2015.2921.

41.

Son

Kim

Cho

, et al. Promoter hypermethylation of the CFTR gene and clinical/pathological features associated with non-small cell lung cancer. Respirology. 2011;16(8):1203-1209. doi:10.1111/j.1440-1843.2011.01994.x.

42.

Haque

Cajas-Salazar

, et al. CYP2E1 polymorphism, cigarette smoking, p53 expression, and survival in non-small cell lung cancer: a long term follow-up study. Appl Immunohistochem Mol Morphol. 2004;12(4):315-322. doi:10.1097/00129039-200412000-00005.

43.

Xiao

Wang

, et al. HDGF-related protein-3 is required for anchorage-independent survival and chemoresistance in hepatocellular carcinomas. Gut. 2013;62(3):440-451. doi:10.1136/gutjnl-2011-300781.

44.

Ocak

Pedchenko

Chen

, et al. Loss of polymeric immunoglobulin receptor expression is associated with lung tumourigenesis. Eur Respir J. 2012;39(5):1171-1180. doi:10.1183/09031936.00184410.

45.

Shen

Cao

Jia

Angiogenesis in liver metastasis of colo-rectal carcinoma. Front Biosci (Landmark Ed). 2013;18:1435-1443. doi:10.2741/4190.

46.

Liu

Mai

Yang

, et al. Candidate tumour suppressor CCDC19 regulates miR-184 direct targeting of C-Myc thereby suppressing cell growth in non-small cell lung cancers. J Cell Mol Med. 2014;18(8):1667-1679. doi:10.1111/jcmm.12317.

47.

Leithner

Wohlkoenig

Stacher

, et al. Hypoxia increases membrane metallo—endopeptidase expression in a novel lung cancer ex vivo model—role of tumor stroma cells. BMC Cancer. 2014;14:40. doi:10.1186/1471-2407-14-40.

48.

Goldstein

Trivedi

Speth

RC.

Alterations in gene expression of components of the renin-angiotensin system and its related enzymes in lung cancer. Lung Cancer Int. 2017;2017:6914976. doi:10.1155/2017/6914976.

49.

Werner

Stamm

Pandjaitan

, et al. Iroquois homeobox 2 suppresses cellular motility and chemokine expression in breast cancer cells. BMC Cancer. 2015;15:896. doi:10.1186/s12885-015-1907-4.

50.

Froehlich

Versapuech

Megrelis

, et al. FAM65B controls the proliferation of transformed and primary T cells. Oncotarget. 2016;7(39):63215-63225. doi:10.18632/oncotarget.11438.

51.

Sun

Liu

, et al. Rig-G is a growth inhibitory factor of lung cancer cells that suppresses STAT3 and NF-kappaB. Oncotarget. 2016;7(40):66032-66050. doi:10.18632/oncotarget.11797.

52.

de Wit

Rijntjes

Diepstra

, et al. Analysis of differential gene expression in human melanocytic tumour lesions by custom made oligonucleotide arrays. Br J Cancer. 2005;92(12):2249-2261. doi:10.1038/sj.bjc.6602612.

53.

Jamaati

Khosravi

Abedini

, et al. Three markers in cancerous and healthy cells of patients with non-small-cell lung carcinoma (NSCLC). Asian Pac J Cancer Prev. 2019;20(8):2281-2285. doi:10.31557/APJCP.2019.20.8.2281.

54.

Bergethon

Shaw

, et al. ROS1 rearrangements define a unique molecular class of lung cancers. J Clin Oncol. 2012;30(8):863-870. doi:10.1200/JCO.2011.35.6345.

55.

Fan

Shen

Yuan

YF.

Expression and prognostic roles of TRPV5 and TRPV6 in non-small cell lung cancer after curative resection. Asian Pac J Cancer Prev. 2014;15(6):2559-2563. doi:10.7314/apjcp.2014.15.6.2559.

56.

Huang

Wei

, et al. SOX9 drives the epithelial-mesenchymal transition in non-small-cell lung cancer through the Wnt/beta-catenin pathway. J Transl Med. 2019;17(1):143. doi:10.1186/s12967-019-1895-2.

57.

Zhou

, et al. Clinical significance of SOX9 in human non-small cell lung cancer progression and overall patient survival. J Exp Clin Cancer Res. 2012;31:18. doi:10.1186/1756-9966-31-18.

58.

Miyamori

Takino

Kobayashi

, et al. Claudin promotes activation of pro-matrix metalloproteinase-2 mediated by membrane-type matrix metalloproteinases. J Biol Chem. 2001;276(30):28204-28211. doi:10.1074/jbc.M103083200.

59.

Urosevic

Garcia-Albéniz

Planet

, et al. Colon cancer cells colonize the lung from established liver metastases through p38 MAPK signalling and PTHLH. Nat Cell Biol. 2014;16(7):685-694. doi:10.1038/ncb2977.

60.

Katono

Sato

Kobayashi

, et al. Clinicopathological significance of S100A14 expression in lung adenocarcinoma. Oncol Res Treat. 2017;40(10):594-602. doi:10.1159/000478100.

61.

Zhu

Xiong

Chen

Lin

Role of high expression levels of STK39 in the growth, migration and invasion of non-small cell type lung cancer cells. Oncotarget. 2016;7(38):61366-61377. doi:10.18632/oncotarget.11351.

62.

Meng

Zhi

Chao

, et al. Bex2 controls proliferation of human glioblastoma cells through NF-kappaB signaling pathway. J Mol Neurosci. 2014;53(2):262-270. doi:10.1007/s12031-013-0215-1.

63.

Xiao

Chen

, et al. BEX2 promotes tumor proliferation in colorectal cancer. Int J Biol Sci. 2017;13(3):286-294. doi:10.7150/ijbs.15171.

64.

Sasahira

Nishiguchi

Kurihara-Shimomura

Nakashima

Kuniyasu

Kirita

NIPA-like domain containing 1 is a novel tumor-promoting factor in oral squamous cell carcinoma. J Cancer Res Clin Oncol. 2018;144(5):875-882. doi:10.1007/s00432-018-2612-x.

65.

Zhang

Dong

, et al. Osteoglycin (OGN) inhibits cell proliferation and invasiveness in breast cancer via PI3K/Akt/mTOR signaling pathway. Onco Targets Ther. 2019;12:10639-10650. doi:10.2147/OTT.S222967.

66.

Peng

Cai

SJ.

Osteoglycin (OGN) reverses epithelial to mesenchymal transition and invasiveness in colorectal cancer via EGFR/Akt pathway. J Exp Clin Cancer Res. 2018;37(1):41. doi:10.1186/s13046-018-0718-2.

67.

Yang

Song

Chen

, et al. Integrated analysis of microfibrillar-associated proteins reveals MFAP4 as a novel biomarker in human cancers. Epigenomics. 2019;11(1):1635-1651. doi:10.2217/epi-2018-0080.

68.

Qiu

, et al. Differentially expressed protein-coding genes and long noncoding RNA in early-stage lung cancer. Tumour Biol. 2015;36(12):9969-9978. doi:10.1007/s13277-015-3714-6.

69.

Han

Wang

Han

, et al. Low-expression of TMEM100 is associated with poor prognosis in non-small-cell lung cancer. Am J Transl Res. 2017;9(5):2567-2578.

70.

Czerwinski

McLemore

Gelboin

Gonzalez

FJ.

Quantification of CYP2B7, CYP4B1, and CYPOR messenger RNAs in normal human lung and lung tumors. Cancer Res. 1994;54(4):1085-1091.

71.

Wang

Chen

, et al. A tropomyosin receptor kinase family protein, NTRK2 is a potential predictive biomarker for lung adenocarcinoma. PeerJ. 2019;7:e7125. doi:10.7717/peerj.7125.

72.

Feng

, et al. DNA methylation-mediated silencing of matricellular protein dermatopontin promotes hepatocellular carcinoma metastasis by alpha3beta1 integrin-Rho GTPase signaling. Oncotarget. 2014;5(16):6701-6715. doi:10.18632/oncotarget.2239.

73.

Tian

Zhao

Fan

Kang

Weighted gene co-expression network analysis in identification of metastasis-related genes of lung squamous cell carcinoma based on the Cancer Genome Atlas database. J Thorac Dis. 2017;9(1):42-53. doi:10.21037/jtd.2017.01.04.

74.

Ding

Tong

Liu

Moscow

Tai

HH.

NAD+-linked 15-hydroxyprostaglandin dehydrogenase (15-PGDH) behaves as a tumor suppressor in lung cancer. Carcinogenesis. 2005;26(1):65-72. doi:10.1093/carcin/bgh277.

75.

Liu

Ren

Song

Upregulation of SPOCK2 inhibits the invasion and migration of prostate cancer cells by regulating the MT1-MMP/MMP2 pathway. PeerJ. 2019;7:e7163. doi:10.7717/peerj.7163.

76.

Shigematsu

Lin

Takahashi

, et al. Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst. 2005;97(5):339-346. doi:10.1093/jnci/dji055.

77.

Shigematsu

Gazdar

AF.

Somatic mutations of epidermal growth factor receptor signaling pathway in lung cancers. Int J Cancer. 2006;118(2):257-262. doi:10.1002/ijc.21496.

78.

Markman

Javier Ramos

Capdevila

Tabernero

EGFR and KRAS in colorectal cancer. Adv Clin Chem. 2010;51:71-119.

79.

Kosaka

Yatabe

Endoh

Kuwano

Takahashi

Mitsudomi

Mutations of the epidermal growth factor receptor gene in lung cancer: biological and clinical implications. Cancer Res. 2004;64(24):8919-8923. doi:10.1158/0008-5472.CAN-04-2818.

80.

ILSI Risk Science Institute. The relevance of the rat lung response to particle overload for human risk assessment: a workshop consensus report. Inhal Toxicol. 2000;12(1-2):1-17. doi:10.1080/08958370050029725.

81.

Grande

Franzen

Karlsson

Ericson

Heldin

Nilsson

Transforming growth factor-beta and epidermal growth factor synergistically stimulate epithelial to mesenchymal transition (EMT) through a MEK-dependent mechanism in primary cultured pig thyrocytes. J Cell Sci. 2002;115(pt 22):4227-4236. doi:10.1242/jcs.00091.

82.

Xie

Law

Chytil

Brown

Aakre

Moses

. Activation of the Erk pathway is required for TGF-beta1-induced EMT in vitro. Neoplasia. 2004;6(5):603-610. doi:10.1593/neo.04241.

83.

Zavadil

Bitzer

Liang

, et al. Genetic programs of epithelial cell plasticity directed by transforming growth factor-beta. Proc Natl Acad Sci U S A. 2001;98(12):6686-6691. doi:10.1073/pnas.111614398.

84.

Shin

Dimitri

Yoon

Dowdle

Blenis

ERK2 but not ERK1 induces epithelial-to-mesenchymal transformation via DEF motif-dependent signaling events. Mol Cell. 2010;38(1):114-127. doi:10.1016/j.molcel.2010.02.020.

85.

Kou

Kandpal

RP.

Differential expression patterns of Eph receptors and ephrin ligands in human cancers. Biomed Res Int. 2018;2018:7390104. doi:10.1155/2018/7390104.

86.

Brantley-Sieders

DM.

Clinical relevance of Ephs and ephrins in cancer: lessons from breast, colorectal, and lung cancer profiling. Semin Cell Dev Biol. 2012;23(1):102-108. doi:10.1016/j.semcdb.2011.10.014.

87.

Kinch

Moore

Harpole

Jr.

Predictive value of the EphA2 receptor tyrosine kinase in lung cancer recurrence and survival. Clin Cancer Res. 2003;9(2):613-618.

88.

Sun

Jiang

, et al. Knockdown of ephrin receptor A7 suppresses the proliferation and metastasis of A549 human lung cancer cells. Mol Med Rep. 2016;13(4):3190-3196. doi:10.3892/mmr.2016.4904.

89.

Kullander

Klein

Mechanisms and functions of Eph and ephrin signalling. Nat Rev Mol Cell Biol. 2002;3(7):475-486. doi:10.1038/nrm856.

90.

Miao

Wei

Peehl

, et al. Activation of EphA receptor tyrosine kinase inhibits the Ras/MAPK pathway. Nat Cell Biol. 2001;3(5):527-530. doi:10.1038/35074604.

91.

Strassheim

Shafer

Phelps

Williams

CL.

Small cell lung carcinoma exhibits greater phospholipase C-beta1 expression and edelfosine resistance compared with non-small cell lung carcinoma. Cancer Res. 2000;60(10):2730-2736.

92.

Ellenbroek

Collard

JG.

Rho GTPases: functions and association with cancer. Clin Exp Metastasis. 2007;24(8):657-672. doi:10.1007/s10585-007-9119-1.

93.

Fritz

Just

Kaina

Rho GTPases are over-expressed in human tumors. Int J Cancer. 1999;81(5):682-687. doi:10.1002/(sici)1097-0215(19990531)81:5<682::aid-ijc2>3.0.co;2-b.

94.

Zhang

Zhou

, et al. Overexpression of RhoE has a prognostic value in non-small cell lung cancer. Ann Surg Oncol. 2007;14(9):2628-2635. doi:10.1245/s10434-007-9457-x.

95.

Sahai

Olson

Marshall

CJ.

Cross-talk between Ras and Rho signalling pathways in transformation favours proliferation and increased motility. EMBO J. 2001;20(4):755-766. doi:10.1093/emboj/20.4.755.

96.

Schaller

MD.

Paxillin: a focal adhesion-associated adaptor protein. Oncogene. 2001;20(44):6459-6472. doi:10.1038/sj.onc.1204786.

97.

Salgia

Ewaniuk

, et al. Expression of the focal adhesion protein paxillin in lung cancer and its relation to cell motility. Oncogene. 1999;18(1):67-77. doi:10.1038/sj.onc.1202273.

98.

Chen

Zhang

, et al. Overexpression of integrin-linked kinase correlates with malignant phenotype in non-small cell lung cancer and promotes lung cancer cell invasion and migration via regulating epithelial-mesenchymal transition (EMT)-related genes. Acta Histochem. 2013;115(2):128-136. doi:10.1016/j.acthis.2012.05.004.

99.

Karachaliou

Cardona

Bracht

JWP

, et al. Integrin-linked kinase (ILK) and src homology 2 domain-containing phosphatase 2 (SHP2): novel targets in EGFR-mutation positive non-small cell lung cancer (NSCLC). EBioMedicine. 2019;39:207-214. doi:10.1016/j.ebiom.2018.11.036.

100.

Schmit

Michiels

TMEM proteins in cancer: a review. Front Pharmacol. 2018;9:1345. doi:10.3389/fphar.2018.01345.

101.

Liu

Shang

Song

LncRNA MEG8 promotes tumor progression of non-small cell lung cancer via regulating miR-107/CDK6 axis. Anticancer Drugs. 2020;31(10):1065-1073. doi:10.1097/CAD.0000000000000970.

102.

Tasheva

Koester

Paulsen

, et al. Mimecan/osteoglycin-deficient mice have collagen fibril abnormalities. Mol Vis. 2002;8:407-415.

103.

Bauer

Velmurugan

Xiong

Alexander

Xiong

Brooks

Epiregulin is required for lung tumor promotion in a murine two-stage carcinogenesis model. Mol Carcinog. 2017;56(1):94-105. doi:10.1002/mc.22475.

104.

Sunaga

Kaira

Epiregulin as a therapeutic target in non-small-cell lung cancer. Lung Cancer (Auckl). 2015;6:91-98. doi:10.2147/LCTT.S60427.

105.

Busser

Coll

Hurbin

The increasing role of amphiregulin in non-small cell lung cancer. Pathol Biol (Paris). 2009;57(7-8):511-512. doi:10.1016/j.patbio.2008.10.002.

106.

Riva

Pandiri

, et al. The mutational signature profile of known and suspected human carcinogens in mice. Nat Genet. 2020;52(11):1189-1197. doi:10.1038/s41588-020-0692-4.