Hexavalent chromium exposure activates the non-canonical nuclear factor kappa B pathway to promote immune checkpoint protein programmed death-ligand 1 expression and lung carcinogenesis

Abstract

Lung cancer is the leading cause of cancer-related death worldwide; however, the mechanism of lung carcinogenesis has not been clearly defined. Chronic exposure to hexavalent chromium [Cr(VI)], a common environmental and occupational pollutant, causes lung cancer, representing an important lung cancer etiology factor. The mechanism of how chronic Cr(VI) exposure causes lung cancer remains largely unknown. By using cell culture and mouse models and bioinformatics analyses of human lung cancer gene expression profiles, this study investigated the mechanism of Cr(VI)-induced lung carcinogenesis. A new mouse model of Cr(VI)-induced lung carcinogenesis was developed as evidenced by the findings showing that a 16-week Cr(VI) exposure (CaCrO₄ 100 μg per mouse once per week) via oropharyngeal aspiration induced lung adenocarcinomas in male and female A/J mice, whereas none of the sham-exposed control mice had lung tumors. Mechanistic studies revealed that chronic Cr(VI) exposure activated the non-canonical NFκB pathway through the long non-coding RNA (lncRNA) ABHD11-AS1/deubiquitinase USP15-mediated tumor necrosis factor receptor-associated factor 3 (TRAF3) down-regulation. The non-canonical NFκB pathway activation increased the interleukin 6 (IL-6)/Janus kinase (Jak)/signal transducer and activator of transcription 3 (Stat3) signaling. The activation of the IL-6/Jak signaling axis by Cr(VI) exposure not only promoted inflammation but also stabilized the immune checkpoint molecule programmed death-ligand 1 (PD-L1) protein in the lungs, reducing T lymphocyte infiltration to the lungs. Given the well-recognized critical role of PD-L1 in inhibiting anti-tumor immunity, these findings suggested that the lncRNA ABHD11-AS1-mediated non-canonical NFκB pathway activation and PD-L1 up-regulation may play important roles in Cr(VI)-induced lung carcinogenesis.

1. Introduction

Lung cancer is the second most common cancer but the leading cause of cancer-related death. However, the mechanism of how lung cancer develops and progresses has not been well understood. Among many risk factors for lung cancer, exposure to metal carcinogens represents an important lung cancer etiology factor. One of the most common lung cancer-causing metal carcinogens is hexavalent chromium [Cr(VI)]. Cr (VI) is a naturally present metal element in rocks and a significant amount of Cr(VI) has been released into the environment due to natural and human activities. As a result, Cr(VI) is listed as the 17th toxicant of top 20 hazard substances in the Agency for Toxic Substances and Disease Registry’s Substance Priority List [1]. While Cr(VI) has been classified as a Group I carcinogen and exposure to Cr(VI) causes lung cancer in humans, the mechanism of Cr(VI) carcinogenesis has not been well understood. Studies have shown that Cr(VI) exposure causes genotoxic effects, epigenetic and epitranscriptomic dysregulations [2–7], which may play important roles in Cr(VI) carcinogenesis. Studies also showed that Cr(VI) exposure causes inflammatory effects in the lung, however, the mechanisms by which Cr(VI) exposure triggers the inflammatory response and its role in Cr(VI) lung carcinogenesis are not well understood [8].

While acute inflammation plays a critical role in wound repair and fighting against viral and bacterial pathogens, persistent and chronic inflammation has been linked to cancer development and progression [9–12]. The nuclear factor kappa B (NFκB) signaling is considered as a master regulator of inflammatory responses, and NFκB signaling has both pro- and anti-inflammatory roles [13–15]. Basically, the NFκB pathway consists of classical (canonical) and alternative (non-canonical) pathways. The NFκB transcription factor family has 5 members: p65 (RelA), p105/p50 (NFκB1), C-Rel, p100/p52 (NFκB2) and RelB. Among the 5 members, NFκB1 and NFκB2 are produced as p105 and p100 precursors, which are proteolytically processed to produce p50 and p52 transcription factors, respectively. In most cases, the RelA (p65)/p50 heterodimer nuclear localization is the indication of the canonical NFκB pathway activation, while the RelB/p52 heterodimer nuclear localization represents the activation of the non-canonical NFκB pathway. While extensive studies on the role of the canonical NFκB pathway in inflammation and cancer have been done, much less is known about the role of the non-canonical NFκB pathway in inflammation and cancer.

The non-canonical NFκB pathway is activated by the proteolytic processing of NFκB2 (p100) [16–18]. Under an unstimulated situation, p100 interacts with RelB and sequesters RelB in the cytoplasm. The p100 processing is mediated by phosphorylation of p100 b y the IκB kinase α (IKKα) homodimer. The kinase activity of IKKα is activated by an up-stream kinase, NF-κB-inducing kinase (NIK), which is stabilized/accumulated upon the proteasome degradation of components in its inhibitory complex consisting of tumor necrosis factor receptor-associated factor 3 (TRAF3), TRAF2, and E3 ligases cellular inhibitor of apoptosis 1 (cIAP1) and cIAP2 [16–18]. Under the basal condition, the ubiquitin ligase cIAP1 and cIAP2 in the inhibitory complex ubiquitinate NIK for degradation. Under the stimulated condition, cIAP1 and cIAP2 ubiquitinate TRAF3 for degradation releasing NIK from the inhibitory complex, thus leading to NIK stabilization and accumulation [18]. Studies showed that TRAF3 is a key inhibitor of the non-canonical NFκB pathway, as knocking down or knocking out TRAF3 is sufficient to induce NIK stabilization, IKKα activation and subsequent p100 processing [19,20]. The proteolytic processing of p100 produces p52 and at the same time, frees RelB, leading to p52/RelB dimer formation, nuclear localization and thus activation of the non-canonical NFκB pathway. While studies have shown that metal carcinogen exposure activates the canonical NFκB pathway, the effect of metal carcinogen exposure on non-canonical NFκB signaling and its role in metal carcinogenesis has been rarely studied.

In this study, we generated a new mouse model to study the mechanism of the carcinogenic effects of chronic Cr(VI) exposure to the lungs. Our mechanistic study revealed that chronic Cr(VI) exposure increased TRAF3 proteolytic degradation to activate the non-canonical NFκB pathway, which increases the expression of interleukin-6 (IL-6), leading to the activation of the Janus kinases (Jaks) and subsequent stabilization of the immune checkpoint protein programmed death-ligand 1 (PD-L1). Given the well-recognized critical role of PD-L1 in inhibiting anti-tumor immunity, these findings suggested that non-canonical NFκB pathway activation and PD-L1 up-regulation may play important roles in Cr(VI)-induced lung carcinogenesis.

2. Materials and methods

2.1. Cell lines and reagents

Immortalized non-tumorigenic human bronchial epithelial BEAS-2B cells and human lung cancer cell lines were purchased from the America Type Culture Collection (ATCC, Manassas, VA). Passage-matched control BEAS-2B cells (BEAS-2B-Control) and a low dose Cr(VI) (K₂Cr₂O₇) exposure-transformed cells [BEAS-2B–Cr(VI)] were generated, characterized and reported in our previous publication [21]. The detailed procedures for generating and verifying ABHD11-AS1, ABHD11-AS1 antisense RNA (ABHD11-AS1-as), or USP15 stably expressing cells or SART3, ABHD11-AS1 stable knockdown cells were described in our recent paper [22]. BEAS-2B cells and human lung adenocarcinoma A549 cells were cultured in Gibco^™ Dulbecco’s Modified Eagle Medium (DMEM) (Thermo Fisher, Waltham, MA) supplemented with 5% fetal bovine serum (FBS) (Sigma-Aldrich, St. Louis, MO) and 1% Penicillin-Streptomycin (P/S) (10,000U/mL) (Thermo Fisher). Human lung adenocarcinoma H460 cells were cultured in Gibco^™ Rosewell Park Memorial Institute (RPMI) 1640 (Thermo Fisher) media, supplemented with 5% FBS and 1% P/S. All cells were maintained at 37 °C in a humidified atmosphere of 5% CO₂ and sub-cultured every 2–3 days using Gibco^™ 0.25% Trypsin-EDTA (Thermo Fisher). Cycloheximide (CHX) and (S)-MG132 were purchased from Cayman Chemical (Ann Arbor, MI). Ruxolitinib was purchased from Advanced ChemBlocks (Hayward, CA). Dharmacon^™ ON-TARGETplus SMARTpool siRNA oligoes specifically targeting human RelB, TRAF3 or USP15 genes, as well as non-targeting control siRNAs, were purchased from Horizon Discovery (Lafayette, CO). Recombinant human IL-6 protein was purchased from R&D Systems (Minneapolis, MN).

2.2. Mouse calcium chromate exposure via oropharyngeal aspiration

Five-week-old male and female A/J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). After one week of acclimation, mice were randomly divided into two groups: the sham exposure control group and the calcium chromate exposure group. Each group contained 15 males and 15 females. Mouse sham or calcium chromate (CaCrO₄, 100 μg per mouse per week) exposure starting at the age of six-week-old was carried out via oropharyngeal aspiration, once per week for 16 weeks. The details of the oropharyngeal aspiration procedure, justification for the dose of calcium chromate and the characterization of the particulate calcium chromate were described in our previous publication [23]. Our animal protocol was reviewed and approved by the Institutional Animal Care and Use Committee of Stony Brook University. The sham control group mice were administered phosphate-buffered saline (PBS). One week after the final oropharyngeal aspiration procedure, all mice were euthanized. The lungs from five male and female mice were harvested and immediately frozen either for RNA extraction, for single cell analysis, or for fixation in 10% neutral-buffered formalin solution for paraffin-embedding and sectioning, followed by histologic processing, hematoxylin and eosin (H&E) staining, and immunohistochemistry (IHC).

2.3. RNA isolation and quantitative real-time PCR analysis

Total RNA was extracted by using the Direct-zol RNA Miniprep Kit (Zymo Research, Irvine, CA) with TRIzol reagent (Thermo Fisher), following the manufacturer’s instructions and RNA concentrations were measured using a NanoDrop spectrophotometer (Thermo Fisher). TaqMan^™ Fast Advanced MasterMix and qPCR probes were added into a MicroAmp^™ Optical 96-Well Reaction Plate for TaqMan Gene Expression Assays. QuantStudio^™ 3 Real-Time PCR System and QuantStudio^™ Data Analysis were performed using the 2^−ΔΔct method to determine the relative expression level of each gene, with human gene RN18S1 and mouse gene Rn18s as the internal controls.

2.4. Subcellular protein fractionation and western blot analysis

Cell pellets were resuspended and lysed with Cytoplasmic Extraction Reagent (CER) and Nuclear Extraction Reagent (NER) (Thermo Fisher) following the manufacturer’s instructions. Briefly, pellets were incubated with ice-cold CER Buffer I on ice for 10 min after vortex vigorously for 10 s. The cell suspension was mixed with CER Buffer II, spun down at 12,000 rpm at 4 °C for 5 min and the cytoplasmic extract was collected separately. Nuclear pellets were further incubated with NER Buffer containing protease inhibitor prior to sonication. For other protein samples extracted from whole cells, pellets were completely lysed by a modified lysis buffer according to our recent article [22]. Western blot analyses were performed following the protocol described in our recent publication [22]. Briefly, all protein samples were applied to Bio-Rad DC^™ Protein Assay (Bio-Rad, Hercules, CA), to determine protein concentrations before SDS-Polyacrylamide Gel Electrophoresis. The separated proteins were transferred to the Immobilon^®-FL PVDF Membrane (Millipore, Burlington, MA). The 5% non-fat dry milk (LabScientific, Danvers, MA) in 1X PBS-buffered saline with Tween^® 20 (PBST) was used for membrane blocking before primary antibody incubation. The primary antibodies and dilution concentration are as follows: anti-GAPDH (#5174, 1:1000), anti–NF–κB p65 (#8242, 1:1000), anti-PD-L1 (#13684, 1:1000), anti-RelB (#10544, 1:1000), anti-Stat3 (#30835, 1:1000), anti-phospho-Stat3 (Tyr705) (#9145, 1:1000), anti-TRAF2 (#4724, 1:1000), anti-TRAF3 (#4729, 1:1000), anti-Ubiquitin (#3936, 1:1000) from Cell Signaling Technology (CST, Danvers, MA); anti-Lamin A/C (sc-7292, 1:250), anti-USP15 (sc-100629, 1:250) from Santa Cruz Biotechnology (SCBT, Dallas, TX); anti-β-actin (A5441, 1: 8000) from Sigma-Aldrich.

2.5. Immunoprecipitation (IP) and co-immunoprecipitation (co-IP) analyses

Cell pellets were lysed on ice for 10 min in 20 mM Tris HCl, 100 mM NaCl, 0.5% NP-40, 0.5 mM EDTA, 1 mM PMSF and protease inhibitors, and immunoprecipitation (IP) was carried out following the protocol described in our previous publication [24]. Briefly, cell suspensions were centrifuged at 13,000 rpm at 4 °C for 10 min. To minimize non-specific binding, one mg of protein lysate was pre-cleared by incubation with IgG and Protein A/G PLUS-Agarose (SCBT) for 2 h. Following this, the supernatant was incubated with 4 μg of primary antibodies (anti-TRAF2, anti-TRAF3 or anti-USP15) at 4 °C overnight. Mouse IgG1 Isotype was used as the negative control. On the following day, Protein A/G PLUS-Agarose (50 μL/sample) was added and co-incubated with supernatant at 4 °C for 2 h before spinning down and stringently washing the mixture. The protein precipitates were further used for standard Western blot analysis. The primary antibodies and dilution concentration are as below: Mouse IgG1 Isotype Control (#5415, 2 μg/μL) from CST; anti-TRAF2 (sc-136999, 1:100), anti-TRAF3 (sc-6933, 1:50), anti-USP15 (sc-515688, 1:100) from SCBT.

2.6. Immunofluorescence stain

Cells were seeded on microscope square coverslips in 6-well plates and incubated at 37 °C for indicated times. Briefly, cells were washed with 1X PBS twice, fixed with 4% paraformaldehyde for 15 min at room temperature (RT), and permeabilized with 1% Triton X-100 for 15 min prior to blocking with 3% bovine serum albumin (BSA) for 30 min. Next, anti–NF–κB p65 (1:750) or anti-RelB (1:1500) diluted in 1% BSA were incubated with cells at 4 °C overnight, followed by incubating with Alexa Fluor^® 546 Goat anti-rabbit IgG (1:250) at RT for 1 h. At the end of incubation, cells were washed twice with PBS and stained with DAPI (4,6-diamidino-2-phenylindole) for 10 min to visualize the nuclei. The images were captured and analyzed using a Leica THUNDER Imaging System (Wetzlar, Germany).

2.7. Immunohistochemistry

Formalin-fixed, paraffin-embedded (FFPE) mouse and human lung tumor tissue sections were treated stepwise with VECTASTAIN Elite ABC Universal PLUS Kit (Vector Laboratories, Newark, CA) according to the manufacturer’s instructions. Briefly, tissue slides were deparaffinized and hydrated through fresh xylenes and graded alcohol series. Antigen retrieval was performed by incubation in an autoclave for 15 min in a citrate-based retrieval solution (pH 6.0) and endogenous peroxidase was quenched by incubation in BLOXALL Blocking Solution for 10 min. Slides were then blocked with 2.5% normal horse serum for 20 min at RT, followed by incubation with the individual primary antibodies at 4 °C overnight. The following day, slides were incubated with prediluted biotinylated horse anti-mouse/rabbit IgG secondary antibody, followed by VECTASTAIN Elite ABC Reagent for 30 min each prior to adding ImmPACT DAB EqV HRP Substrate and counterstaining with 50% Gill’s Hematoxylin staining solution (StatLab, McKinney, TX). The primary antibodies and dilution concentration were as follows: anti-Ki-67 (#9027, 1:500), anti-PD-L1 (#13684, 1:200), anti-RelB (#10544, 1:1500), anti-phospho-Stat3 (Tyr705) (#9145, 1:200) from CST; anti-pan-Cytokeratin (#MA5–13203, 1:200) from Thermo Fisher.

2.8. Flow cytometry analysis of mouse lung tissue single cells

Following a 16-week chronic exposure to either Cr(VI) or PBS via oropharyngeal aspiration in both male and female A/J mice as described above, mice were euthanized and lung tissues were collected, cut into small pieces, and dissociated using the gentleMACS^™ Dissociator from Miltenyi Biotec (Bergisch-Gladbach, Germany). Subsequently, the resulting single cell suspension was filtered through a MACS^® SmartStrainer, followed by thorough washing and centrifugation. To eliminate red blood cells, a 1X Red Blood Cell Lysis Solution was prepared and applied to cell pellets. Cell count and viability were determined using Trypan Blue Solution, and the cells were then frozen and stored at −80 °C. Before staining, cells were thawed, centrifuged, and subjected to a 15-min incubation with FcγRII/III receptor block solutions (#101301, BioLegend, San Diego, CA). After washing, cells were incubated for 30 min with specific primary antibodies in dark on ice. Finally, the samples underwent flow cytometric analysis using a MA900 Multi-Application Cell Sorter (SONY, New York, NY), and the data were further analyzed by FlowJo^™ software (BD, Franklin Lakes, NJ). The following antibodies were employed for flow cytometry: FITC-conjugated anti-CD45 (#103108, BioLegend), APC/Cyanine7-conjugated anti-CD3 (#100222, BioLegend), BV785-conjugated anti-PD-L1 (#124331, BioLegend), and 7-AAD (#4204030, BioLegend).

2.9. ICP-MS analysis of mouse lung tissue chromium levels

Mouse lung tissue total chromium levels were measured using a Nexion 350C ICP-MS. Briefly, freshly frozen mouse lung tissues were thawed, weighed (approximately 25 mg of the chromate-exposed or 60 mg of the PBS-exposed mouse lung tissues) and digested in 7 mL of the Optima grade Nitric Acid (Fisher Scientific). The digested mixtures were heated on a hot plate in a fume hood until approximately 120 μL of acid remained. The digests were transferred into 15 mL acid washed polypropylene tubes containing 2 mL of Milli-Q water. Total volume of each digestion was adjusted to 6 mL with Milli-Q water and analyzed on a Nexion 350C ICP-MS using Kinetic Energy Discrimination (KED-Helium gas mode). Internal standards were added to each sample before measurement to control sample matrix interference. Sample Standard Reference Material^® 1566b Oyster Tissue (NIST) and spike recoveries were used to determine method performance.

2.10. ELISA analysis of IL-6 protein levels in cultured cells and mouse lung tissues

The standard ELISA protocol was used to analyze IL-6 protein levels in cultured cells and mouse lung tissues. The ELISA kits for mouse IL-6 (Cat# 431304) and human IL-6 (Cat# 430504) were purchased from Biolegend (San Diego, CA).

2.11. Bioinformatic analyses

The Excel file showing the differentially-expressed genes between passage-matched control cells (BEAS-2B-Control) and Cr(VI)-transformed cells [BEAS-2B–Cr(VI)] was included in our recent paper [22]. The Excel files showing the differentially-expressed genes between Cr(VI)-transformed cells with a vector control [BEAS-2B–Cr(VI)-Control] and Cr(VI)-transformed cells stably expressing ABHD11-AS1 antisense RNA (ABHD11-AS1-as) [BEAS-2B–Cr(VI)-ABHD11-AS1-as] (B), and between human lung adenocarcinoma H460 cells with a vector control (H460-Control) and H460 cells stably expressing ABHD11-AS1-as (H460-ABHD11-AS1-as) are included as supplementary materials (Supplementary Excel File 1–2) The differentially-expressed gene Excel files of these three pairs of cells were used for analyzing common genes that are up-regulated in Cr(VI)-transformed cells but down-regulated in ABHD11-AS1 antisense RNA stably expressing Cr (VI)-transformed cells and human lung cancer H460 cells. The FunRich3 platform (http://www.funrich.org/) (accessed on November 26, 2022) was used for gene visualizing via volcano plot and gene clustering via Venn diagram. The DAVID platform (https://david.ncifcrf.gov/tools.jsp (accessed on November 21, 2023) was used for above identified common gene function classification via functional annotation clustering [25].

The datasets for analyzing the correlation between RelB and ABHD11-AS1 expression levels were obtained from TCGA_LUAD projects (https://portal.gdc.cancer.gov/projects/TCGA-LUAD). The MedCalc statistical platform (https://www.mdcalc.com/) (accessed on November 26, 2022) was used to analyze the correlation between RelB and ABHD11-AS1 expression levels. The expression profiles of RelB in lung adenocarcinomas (LUADs) and lung squamous cell carcinomas (LUSCs) obtained from TCGA_LUAD (https://portal.gdc.cancer.gov/projects/TCGA-LUAD) and TCGA_LUSC (https://portal.gdc.cancer.gov/projects/TCGA-LUSC) project datasets were used for analyzing RelB expression levels in lung cancer and its prognostic value. MedCalc statistical platform (https://www.mdcalc.com/) (accessed on November 23, 2022) was used to explore the prognostic performance via overall survival analysis of lung cancer patients.

2.12. Statistical analysis

The statistical analyses for the significance of differences in numerical data (mean ± SD) were performed by testing different treatment effects, using two-tailed t tests for comparison of two data sets. A p value of <0.05 was considered statistically significant.

3. Results

3.1. A new mouse model demonstrates the carcinogenic effect of chronic Cr(VI) exposure to the lungs

Early studies showed that pulmonary adenomas, adenocarcinomas and squamous cell carcinomas are observed in rats from nine months to two years after inserting a pellet of particulate chromate (strontium chromate) into the bronchus [26,27]. Our recent study found that massive pulmonary adenomas and adenocarcinomas were detected in a 70-week experiment in male A/J mice exposed to calcium chromate via oropharyngeal aspiration once per week (100 μg of calcium chromate per mouse) for 26 weeks and housed for another 44 weeks in the absence of calcium chromate exposure [23]. However, many sham-exposed control A/J mice also developed spontaneous pulmonary adenomas and adenocarcinomas over the 70-week course of the study [23]. To facilitate the mechanistic study of Cr(VI)-induced pulmonary carcinogenesis, we sought to determine whether lung tumors could only be observed in chromate-exposed A/J mice, but not in sham-exposed control A/J mice over the course of a much shorter experiment period. We performed calcium chromate exposure as described in our previous publication [23] but reduced the time of exposure from 26 weeks to 16 weeks. All mice in both PBS control and chromate exposure groups were still living at the end of the study. All mice were euthanized one week after the last PBS or chromate exposure. At the end of the experiment, the body weight of chromate-exposed mice was about 10% less than that of the PBS control group mice. The ICP-MS analysis of mouse lung tissue chromium levels revealed the chromium accumulation in chromate-exposed mouse lungs, while undetectable or very low levels of chromium were found in PBS-exposed mouse lungs (Supplementary Table 1). The histologic assessment of murine pulmonary tissue histology revealed that all sham PBS-exposed control male or female lungs had normal pulmonary parenchyma and no tumors (Fig. 1A–C). In striking contrast, all chromate-exposed male and female mice developed lung tumors (Fig. 1A–C). There were no significant differences in the incidence of pulmonary tumors or tumor burden between Cr (VI)-exposed male versus female mice (Fig. 1B). Histologic analyses revealed moderately differentiated adenocarcinomas (Fig. 1C). The tumor cells were arranged as poorly defined acinar to duct-like structures with crowded, moderately pleomorphic, oval nuclei (Fig. 1C). Moreover, Chromate-exposed mouse lungs also had mild acute and moderate and focally severe chronic inflammation extending into alveolar septae, in proximity to areas of adenocarcinoma, with scattered alveolar macrophages associated with amorphous foreign bodies and cholesterol clefts (Fig. 1D). The sham PBS-exposed control mice were negative for acute and chronic inflammation. There were no significant differences in chromate exposure-induced lung inflammation and tumorigenesis between male and female mice.

Fig. 1.

A 16-week calcium chromate exposure via oropharyngeal aspiration induces lung tumors in A/J mice. (A) Representative H&E staining whole lung images of mice exposed to either PBS or calcium chromate (100 μg once per week) via oropharyngeal aspiration for 16 weeks. (B) Mouse lung tumor incidence and tumor burden in mice exposed to PBS or calcium chromate for 16-weeks *p < 0.05, **p < 0.01, compared with PBS-exposed mice (Mean ± SD, n = 5). (C) Representative H&E staining high magnification images showing normal pulmonary parenchyma, including benign bronchioles from a mouse exposed to PBS compared to the histology of moderately differentiated adenocarcinoma from a mouse exposed to calcium chromate (100 μg once per week) for 16 weeks. (D) A representative H&E staining high magnification image showing moderately differentiated adenocarcinoma with few macrophages from a mouse exposed to calcium chromate (100 μg once per week) for 16 weeks. Red arrows point to representative macrophages. (E–F) Representative IHC staining for Ki67 (E) and pan-cytokeratin (F) in the lungs of mice exposed to either PBS or calcium chromate (100 μg once per week) for 16 weeks.

Further immunohistochemistry (IHC) analyses showed that chromate-exposed mouse lung tumor areas displayed strong Ki-67 positive nuclear staining (Fig. 1E) in pan-cytokeratin-positive tumor cells, reflecting active cell proliferation (Fig. 1F), confirming that the proliferating cells were mostly epithelial tumor cells. Together, these results indicated that exposure to calcium chromate (100 μg) via oropharyngeal aspiration once a week for 16 weeks induced lung adenocarcinomas in A/J mice.

3.2. Chronic Cr(VI) exposure activates the non-canonical NFκB pathway in cultured cells, mouse and human lung tissues

Our recent study showed that chronic Cr(VI) exposure up-regulates the expression of a long non-coding RNA ABHD11-AS1, which plays an important role in Cr(VI) carcinogenesis and lung cancer [22]. Our mechanistic studies revealed that ABHD11-AS1 interacts with SART3 (spliceosome associated factor 3, U4/U6 recycling protein) to promote a deubiquitinase USP15 (ubiquitin specific peptidase 15) translocation from cytoplasm to nuclear, regulating RNA alternative splicing and enhancing cancer stemness [22]. Since USP15 is an important deubiquitinase and regulates oncogenic signaling pathways that play critical roles in cancer [28–30], we sought to investigate what oncogenic signaling pathway is activated by chronic Cr(VI) exposure-caused ABHD11-AS1 up-regulation and USP15 re-distribution. After analyzing our microarray and RNA-seq data, the differentially-expressed genes were obtained for the following three pairs of cells: (i) passage-matched control human bronchial epithelial BEAS-2B cells (BEAS-2B-Control) and chronic Cr(VI) exposure-transformed cells [BEAS-2B–Cr(VI)] [22]; (ii) Cr(VI)-transformed cells with a vector control [BEAS-2B–Cr (VI)-Control] and Cr(VI)-transformed cells stably expressing ABHD11-AS1 antisense RNA (ABHD11-AS1-as) [BEAS-2B–Cr(VI)-ABHD11-AS1-as] (Supplementary Excel file 1); and (iii) human lung adenocarcinoma H460 cells with a vector control (H460-Control) and H460 cells stably expressing ABHD11-AS1-as (H460-ABHD11-AS1-as) (Supplementary Excel file 2). Thirty-one common genes that were up-regulated in Cr(VI)-transformed cells but down-regulated in ABHD11-AS1 antisense RNA stably expressing Cr(VI)-transformed cells and human lung cancer H460 cells were identified (Supplementary Figs. 1A–E). We performed signaling pathway analysis and found that ten genes out of the thirty-one common genes were involved in the ubiquitin-like modifier conjugation pathway (Supplementary Fig. 2A). Further analysis revealed that three genes (TNFAIP3, BIRC3 and NFKB2) out of these 31 genes and two genes out of the 10 genes were involved in the NFκB signaling pathway (Supplementary Fig. 2B), suggesting that the NFκB signaling pathway may be affected by Cr(VI) exposure possibly through the protein ubiquitination-related mechanism.

We next determined both canonical and non-canonical NFκB pathway activation status in Cr(VI)-transformed cells using both cell fractionation assay and immunofluorescence (IF) staining. As shown in Fig. 2A, the nuclear fractions of p65 and RelB were increased in Cr(VI)-transformed cells compared to control cells. These observations were further confirmed by IF staining, showing much stronger nuclear staining for p65 and RelB in Cr(VI)-transformed cells than the control cells (Fig. 2B). Together, these results suggested that both canonical and non-canonical NFκB pathways are activated in Cr(VI)-transformed cells. We then further determined whether the up-regulation of the lncRNA ABHD11-AS1 plays an important role in both canonical and non-canonical NFκB pathway activations in Cr(VI)-transformed cells. It was found that stably expressing the antisense RNA of ABHD11-AS1 to inhibit ABHD11-AS1 function significantly reduced the nuclear fraction of and nuclear positive staining of RelB in Cr(VI)-transformed cells (Fig. 2C and D). Stably expressing ABHD11-AS1 antisense RNA also reduced p65 nuclear localization in Cr(VI)-transformed cells but to a much less extent (Fig. 2C and D). These results suggested that lncRNA ABHD11-AS1 plays a more important role in chronic Cr(VI) exposure-caused non-canonical NFκB pathway activation than the canonical NFκB pathway activation. Since much less is known about the role of non-canonical NFκB pathway in carcinogenesis, we then focused on the role and mechanism of ABHD11-AS1-mediated non-canonical NFκB pathway activation in Cr(VI) carcinogenesis.

Fig. 2.

Chronic Cr(VI) exposure activates non-canonical NF-κB pathway in human bronchial epithelial cells and mouse lung tissues. (A, C) Representative Western blot images of cell fractionation analysis of p65 and RelB in passage-matched control and chronic Cr(VI) exposure-transformed BEAS-2B cells (A), and in Cr(VI)-transformed BEAS-2B cells with an empty vector or overexpressing the antisense RNA of ABHD11-AS1 (C). The numbers underneath each protein band are the ratios of the intensities of the protein band divided by the corresponding loading control protein band intensities. (B, D) Representative overlaid images of immunofluorescence (IF) staining of RelB or p65 (red) and DNA DAPI (blue) staining in passage-matched control and chronic Cr(VI) exposure-transformed BEAS-2B cells (B), and in Cr(VI)-transformed BEAS-2B cells with an empty vector or overexpressing the antisense RNA of ABHD11-AS1 (D). Scale bar: 50 μm. (E, F) Representative IHC staining of RelB in Cr(VI)-exposed murine (E) or human lungs (F). The murine lung tumor tissue and the adjacent normal lung tissue were from a mouse exposed to calcium chromate (100 μg once per week) for 16 weeks. The human lung tumor tissue and the adjacent normal lung tissue were from a 69-year-old male chromate worker exposed to chromate for 11 years.

We performed IHC of RelB in chronic Cr(VI) exposure-induced mouse and human lung tumor tissues and found that nuclear RelB staining was significantly stronger in chronic Cr(VI) exposure-induced mouse and human lung tumor tissues than the corresponding adjacent normal lung tissues (Fig. 2E and F). Similar RelB staining patterns were also observed in another two Cr(VI)-exposed mice and in the pulmonary tissue of Cr(VI) workers. Moreover, RelB protein levels were higher in both the cytoplasm and nucleus of Cr(VI)-transformed cells (Fig. 2A and B), indicating that chronic Cr(VI) exposure not only activates RelB but also up-regulates RelB expression. Further bioinformatics analyses of human lung cancer gene expression data showed that RelB expression was significantly higher in pulmonary adenocarcinomas and was positively correlated with ABHD11-AS1 expression in lung adenocarcinomas (Supplementary Figs. 3A–B). Moreover, the high expression of RelB was associated with significantly worse overall survival of patients with lung adenocarcinoma, but not in patients with pulmonary squamous cell carcinoma (Supplementary Fig. 3C). These results were consistent with our recent study which reported that ABHD11-AS1 was highly expressed in pulmonary adenocarcinomas and was associated with significantly worse overall survival of patients with pulmonary adenocarcinoma but not pulmonary squamous cell carcinoma [22]. Collectively, these results suggested that ABHD11-AS1 up-regulation may plays an important role in non-canonical NFκB pathway activation by chronic Cr(VI) exposure and that non-canonical NFκB pathway activation may play an important role in Cr(VI) carcinogenesis.

3.3. Chronic Cr(VI) exposure activates the non-canonical NFκB pathway by increasing TRAF3 proteasome degradation

We next sought to further determine the mechanism of chronic Cr (VI) exposure-caused non-canonical NFκB pathway activation. Previous studies showed that TRAF3 is a key inhibitor of the non-canonical NFκB pathway as TRAF3 down-regulation is sufficient to activate the non-canonical NFκB pathway [18–20]. To demonstrate the role of the lncRNA ABHD11-AS1 up-regulation in non-canonical NFκB pathway activation, we first analyzed TRAF3 protein levels in ABHD11-AS1 overexpression human bronchial epithelial cells, ABHD11-AS1 antisense RNA overexpression Cr(VI)-transformed cells and human lung cancer cells. It was found that overexpressing ABHD11-AS1 reduced TRAF3 protein levels in human bronchial epithelial cells (Fig. 3A). In contrast, overexpressing the antisense RNA of ABHD11-AS1 in Cr (VI)-transformed cells and human lung cancer cells increased TRAF3 protein levels (Fig. 3B). These results were consistent with the results presented in Fig. 2C and D showing that stably expressing ABHD11-AS1 antisense RNA reduced non-canonical NFκB pathway activation in Cr (VI)-transformed cells. Moreover, our recent study showed that ABHD11-AS1 functions through interaction with SART3 and knockdown of SART3 reduced USP15 nuclear localization in Cr(VI)-transformed cells and human lung cancer cells [22]. It was further determined that SART3 knockdown reduced TRAF3 ubiquitination and increased TRAF3 protein levels in lung cancer cells (Fig. 3C and D). Together, these results suggested that up-regulation of ABHD11-AS1 activated the non-canonical NFκB pathway possibly through down-regulating TRAF3 protein level by increasing TRAF3 ubiquitination and proteasome degradation.

Fig. 3.

Overexpressing ABHD11-AS1 or its antisense RNA reduces and increases TRAF3 protein levels, respectively. (A, B) Western blot analysis of the effect of overexpressing ABHD11-AS1 in BEAS-2B cells (A) or overexpressing the antisense RNA of ABHD11-AS1 in Cr(VI)-transformed cells and human lung cancer cells (B) on TRAF3 protein levels. The numbers underneath the TRAF3 protein band are the ratios of the intensities of the TRAF3 protein band divided by the corresponding loading control β-actin protein band intensities. (C, D) Immunoprecipitation (IP) and Western blot analysis of the effect of shRNA knockdown of SART3 on TRAF3 ubiquitination level (C) and total protein level (D) in human cancer cells. (E, F) IP analysis of the effect of overexpressing ABHD11-AS1 (E) or the antisense RNA of ABHD11-AS1 (F) on TRAF3 ubiquitination levels. (G, H) IP analysis of the effect of siRNA knocking down USP15 (G) or overexpressing USP15 (H) on TRAF3 ubiquitination levels in ABHD11-AS1 antisense RNA overexpressing human lung cancer cells.

To confirm this point, we next performed TRAF3 immunoprecipitation experiments to detect TRAF3 protein ubiquitination levels. It was found that stably expressing ABHD11-AS1 in human bronchial epithelial cells increased TRAF3 ubiquitination level (Fig. 3E). In contrast, stably expressing the antisense RNA of ABHD11-AS1 reduced TRAF3 ubiquitination level (Fig. 3F). These findings were consistent with the results presented in Fig. 3A and B showing that stably expressing ABHD11-AS1 reduced TRAF3 protein level whereas stably expressing ABHD11-AS1 antisense RNA increased TRAF3 protein level. Our recent study showed that ABHD11-AS1 up-regulation caused the deubiquitinase USP15 translocation from cytoplasm to the nucleus [22], we next determined the role of USP15 in TRAF3 ubiquitination. It was found that knocking down USP15 reversed the effect of stably expressing ABHD11-AS1 antisense RNA on TRAF3 ubiquitination (Fig. 3G). In contrast, overexpressing USP15 further enhanced the effect of stably expressing ABHD11-AS1 antisense RNA on TRAF3 ubiquitination (Fig. 3H). These results suggested that up-regulation of ABHD11-AS1 down-regulates TRAF3 protein level through increasing TRAF3 ubiquitination and proteasome degradation by changing USP15 cellular localization.

We next further determined whether there is a direct interaction between USP15 and TRAF3 proteins. Our co-immunoprecipitation experiments showed that TRAF3 not only directly interacted with TRAF2 (as previously reported) but also interacted with USP15 (Fig. 4A). These results along with the results presented in Fig. 3G and H indicated that the interaction between the deubiquitinase USP15 and TRAF3 can stabilize TRAF3 protein by reducing TRAF3 protein ubiquitination levels, leading to the inhibition of the non-canonical NFκB pathway. This point was further supported by the IF staining of RelB showing that knocking down USP15 or TRAF3 in ABHD11-AS1 antisense RNA overexpressing cells drastically increased RelB nuclear localization and RelB protein levels (Fig. 4B–G), an indication of non-canonical NFκB pathway activation. In addition, the increases of the protein levels of p52 that forms p52/RelB heterodimer provides additional evidence demonstrating non-canonical NFκB pathway activation (Fig. 4C–E, G).

Fig. 4.

USP15 interacts with TRAF3, and knocking down USP15 or TRAF3 reverses the inhibitory effect of overexpressing ABHD11-AS1 antisense RNA on non-canonical NF-kB pathway activation. (A) Representative Western blot images of IP analysis of the interaction between TRAF2 and TRAF3, and USP15 and TRAF3. (B, D, F) Representative overlaid images of IF staining of RelB (red) and DNA DAPI (blue) staining in ABHD11-AS1 antisense RNA overexpressing lung cancer cells (B, F) and Cr(VI)-transformed cells (D) transfected with siRNA control (siCtrl) or siRNA targeting USP15 or TRAF3. Scale bar: 50 μm. (C, E, G) Representative Western blot analysis of the effect of siRNA knocking down USP15 or TRAF3 on RelB and p52 protein levels in ABHD11-AS1 antisense RNA overexpressing lung cancer cells (C, G) and Cr(VI)-transformed cells (E). The numbers underneath each protein band are the ratios of the intensities of the protein bands divided by the corresponding loading control β-actin protein band intensities.

3.4. Non-canonical NFκB pathway activation up-regulates PD-L1 expression level by increasing PD-L1 protein stability

We next investigated the potential role of non-canonical NFκB pathway activation in Cr(VI) carcinogenesis. Given the importance of immune evasion/suppression in carcinogenesis and recent findings showing that high RelB expression level is positively correlated with high expression levels of immune checkpoint molecule PD-L1 in multiple types of cancer including lung cancer [31], we decided to examine whether chronic Cr(VI) exposure up-regulates PD-L1 level and the role of RelB in regulating PD-L1 expression level. It was found that PD-L1 protein level was significantly higher in Cr(VI)-transformed cells than the passage-matched control cells (Fig. 5A). Moreover, stably expressing ABHD11-AS1 alone significantly increased PD-L1 protein levels in human bronchial epithelial cells (Fig. 5A). In contrast, stably expressing the antisense RNA of ABHD11-AS1 significantly reduced the protein level of PD-L1 in human lung cancer cells (Fig. 5A). It was further determined that PD-L1 protein levels were also significantly higher in chronic Cr(VI) exposure-induced mouse and human lung tumor tissues than the corresponding adjacent normal lung tissues (Fig. 5B and C). Similar PD-L1 staining patterns were also observed in another two Cr (VI)-exposed mice and workers’ lung tissues. Together, these results indicated that chronic Cr(VI) exposure increased PD-L1 expression levels in cultured cells and mouse and human lung tissues possibly through ABHD11-AS1 up-regulation.

Fig. 5.

Chronic Cr(VI) exposure increases PD-L1 protein levels in human bronchial epithelial cells, mouse and human lung tumor tissues. (A) Representative Western blot analysis of PD-L1 protein levels in passage-matched control and Cr(VI)-transformed BEAS-2B cells, vector control and ABHD11-AS1 overexpressing BEAS-2B cells, vector control and ABHD11-AS1 antisense RNA overexpressing H460 cells. (B, C) Representative images of PD-L1 IHC staining in Cr(VI)-exposed mouse (B) and human (C) lung tumor and adjacent normal lung tissues. The human lung tissue was from a 69-year-old male chromate worker exposed to chromate for 11 years. (D) Q-PCR analysis of CD274 mRNA levels in passage-matched control and Cr(VI)-transformed BEAS-2B cells, vector control and ABHD11-AS1 overexpressing BEAS-2B cells (Mean ± SD, n = 3). (E) Q-PCR analysis of RelB and CD274 mRNA levels in Cr(VI)-transformed cells and lung cancer H460 cells transfected with control siRNA or RelB targeting siRNA oligoes. *p < 0.05, compared with control siRNA group (Mean ± SD, n = 3). (F) Representative Western blot analysis of RelB and PD-L1 protein levels in Cr(VI)-transformed cells and lung cancer H460 cells transfected with control siRNA or RelB targeting siRNA oligos. (G) Representative Western blot analysis of TRAF3 and PD-L1 protein levels in Cr(VI)-transformed cells and lung cancer H460 cells stably expressing the antisense RNA of ABHD11-AS1 and transfected with control siRNA or TRAF3 targeting siRNA oligos. (H, I) Cycloheximide (CHX) chase assay showing the effect of inhibiting ABHD11-AS1 on PD-L1 protein stability in Cr(VI)-transformed cells. **p < 0.01, compared with ABHD11-AS1 antisense RNA expressing cells at the same time point (Mean ± SD, n = 3). The numbers underneath each protein band are the ratios of the intensities of the protein bands divided by the corresponding loading control β-actin protein band intensities.

We further explored the mechanism of how ABHD11-AS1 up-regulation by chronic Cr(VI) exposure increased PD-L1 expression levels. Interestingly, qPCR analysis showed that chronic Cr(VI) exposure or stably expressing ABHD11-AS1 alone did not significantly increase the mRNA level of CD274 (Fig. 5D), the gene that encodes PD-L1. These results suggested that chronic Cr(VI) exposure increased PD-L1 expression level may occur through posttranscriptional mechanisms. This point is further supported by the results from a loss-of-function approach by stably knocking down ABHD11-AS1 expression in Cr(VI)-transformed cells. It was found that stably knocking down ABHD11-AS1 expression in Cr(VI)-transformed cells greatly reduced their PD-L1 protein levels but had no significant effect on CD274 mRNA levels (Supplementary Fig. 4). All together, these results indicated that ABHD11-AS1 levels were positively correlated with PD-L1 protein levels but were not correlated with CD274 mRNA levels.

To determine the role of RelB in Cr(VI) exposure-caused PD-L1 up-regulation, we performed RelB siRNA knocking down experiments. It was found that RelB targeting siRNAs significantly reduced RelB mRNA levels in Cr(VI)-transformed cells and human lung cancer H460 cells but had no significant effect on CD274 gene mRNA level (Fig. 5E). In contrast, siRNA knocking down RelB drastically reduced the protein levels of PD-L1 in Cr(VI)-transformed cells and human lung cancer cells (Fig. 5F). Consistent with the findings that knocking down TRAF3 in ABHD11-AS1 antisense RNA overexpressing cells reactivated the RelB pathway (Fig. 4B–G), it was found that knocking down TRAF3 greatly increased PD-L1 protein levels in ABHD11-AS1 antisense RNA overexpressing cells (Fig. 5G). These results provided evidence supporting that chronic Cr(VI) exposure up-regulates PD-L1 expression level through ABHD11-AS1-activated RelB via post-transcriptional mechanisms. Indeed, the cycloheximide (CHX) chase experiment showed that stably expressing the antisense RNA of ABHD11-AS1 significantly reduced PD-L1 protein stability in the presence of CHX treatment that inhibited protein translation processes (Fig. 5H). In addition, proteasome degradation inhibitor MG132 treatment drastically recovered PD-L1 protein levels in BEAS-2B control cells and Cr(VI)-transformed cells with ABHD11-AS1 stable knockdown (Supplementary Fig. 5). In contrast, MG132 treatment also increased PD-L1 protein levels in Cr(VI)-transformed cells, ABHD11-AS1 overexpressing cells and Cr(VI)-transformed cells expressing a control shRNA, but to a much less extent (Supplementary Fig. 5). Together, these findings, along with the results presented in Figs. 2 and 3, indicated that chronic Cr(VI) exposure increased PD-L1 expression by regulating PD-L1 protein stability through ABHD11-AS1-mediated non-canonical NFκB pathway activation.

3.5. Non-canonical NFκB pathway activation enhances the IL-6/Jak pathway activity to increase PD-L1 protein stability

We next further determined the mechanism of how non-canonical NFκB pathway activation by chronic Cr(VI) exposure increased PD-L1 protein stability. Previous studies showed that exposure to cigarette smoke increased RelB nuclear localization and its binding to the promoter region of interleukin-6 (IL-6) and other inflammatory genes to increase their expression in mouse lung cells [32,33]. More recent studies showed that the IL-6/Janus kinase 1 (Jak1) pathway activation or Jak2 activation mutation was capable of up-regulating PD-L1 expression by increasing PD-L1 protein stability or CD274 gene transcription [34,35], respectively. We then analyzed IL-6 expression and Jak pathway activity and their roles in PD-L1 up-regulation in Cr (VI)-transformed cells and human lung cancer cells. It was found that chronic Cr(VI) exposure significantly up-regulated IL-6 mRNA and protein levels in cultured human bronchial epithelial cells and mouse lung tissues (Fig. 6A, Supplementary Fig. 6). Stably expressing ABHD11-AS1 alone was capable of significantly increasing IL-6 expression level in human bronchial epithelial cells (Fig. 6B). However, siRNA knocking down RelB significantly reduced IL-6 expression level in Cr(VI)-transformed cells and human lung cancer cells (Fig. 6C). Western blot analysis revealed that phospho-Stat3 (signal transducer and activator of transcription 3) level was drastically higher in Cr (VI)-transformed cells (Fig. 6D). Overexpressing ABHD11-AS1 alone significantly increased phospho-Stat3 levels in human bronchial epithelial cells (Fig. 6E). In contrast, overexpressing the antisense RNA of ABHD11-AS1 greatly reduced phospho-Stat3 levels in Cr (VI)-transformed cells and human lung cancer cells (Fig. 6F). Moreover, IHC staining revealed strong positive phospho-Stat3 staining in chronic Cr(VI) exposure-induced mouse and human lung tumor tissues (Fig. 6G). Similar Phospho-Stat3 staining patterns are also observed in another two Cr(VI)-exposed mice and workers’ lung tissues. Together, these results suggested that chronic Cr(VI) exposure activated the Jak/Stat3 pathway, likely through ABHD11-AS1/RelB-mediated up-regulation of IL-6 expression.

Fig. 6.

Chronic Cr(VI) exposure increases IL-6 expression and activates the Jak/Stat3 pathway. (A, B, C) Q-PCR analysis of IL-6 mRNA levels in passage-matched control and Cr(VI)-transformed cells, sham- and Cr(VI)-exposed mouse lung tissues (A), vector control and ABHD11-AS1 overexpressing cells (B), Cr(VI)-transformed cells and lung cancer H460 cells transfected with control siRNA or RelB targeting siRNA oligos (C). *p < 0.05, **p < 0.01, compared with corresponding control groups (Mean ± SD, n = 3). (D, E, F) Representative Western blot analysis of phospho- and total Stat3 protein levels in passage-matched control and Cr(VI)-transformed cells (D), vector control and ABHD11-AS1 overexpressing cells (E), vector control and ABHD11-AS1 antisense RNA overexpressing Cr(VI)-transformed cells and lung cancer H460 cells (F). (G) Representative images of phospho-Stat3 IHC staining in Cr(VI)-exposed mouse and human lung tumor and adjacent normal lung tissues. The human lung tissue was from a 69-year-old male chromate worker exposed to chromate for 11 years. The numbers underneath each protein band are the ratios of the intensities of the protein bands divided by the corresponding loading control β-actin protein band intensities.

We next determined the role of the IL-6/Jak pathway in the up-regulation of PD-L1 protein levels. Firstly, the Jak inhibitor Ruxolitinib treatment greatly reduced PD-L1 protein levels in Cr(VI)-transformed cells, ABHD11-AS1 overexpressing cells and human lung cancer cells (Fig. 7A), implying an important role of Jak activity in ABHD11-AS1-mediated PD-L1 up-regulation. Secondly, it was found that supplementing recombinant human IL-6 protein was capable of rescuing the effect of RelB knockdown-caused PD-L1 protein decrease in Cr(VI)-transformed cells, ABHD11-AS1 overexpressing cells or human lung cancer cells (Fig. 7B). Collectively, these results indicated that chronic Cr(VI) exposure up-regulates PD-L1 protein levels through the ABHD11-AS1/RelB/IL-6/Jak pathway.

Fig. 7.

Inhibition of Jak activity reduces PD-L1 protein levels. (A) Representative Western blot analysis of the effect of a Jak inhibitor (Ruxolitinib) treatment on PD-L1 protein levels in Cr(VI)-transformed cells, ABHD11-AS1 overexpressing BEAS-2B cells, and human lung cancer H460 cells. (B) Representative Western blot analysis of the effect of RelB knockdown and IL-6 supplementation on phospho-Stat3 and PD-L1 protein levels in Cr(VI)-transformed cells, ABHD11-AS1 overexpressing BEAS-2B cells, and human lung cancer H460 cells. The numbers underneath each protein band are the ratios of the intensities of the protein bands divided by the corresponding loading control β-actin protein band intensities.

3.6. Chronic Cr(VI) exposure reduces T lymphocyte infiltration into lung tissues

We next used flow cytometry analysis to quantitatively determine the effect of chronic Cr(VI) exposure on mouse lung PD-L1 expression at single-cell levels. As shown in Fig. 8A, chronic Cr(VI)-exposed male mouse lung cells had significantly higher PD-L1 levels than the PBS-treated control male mouse lung cells. Chronic Cr(VI)-exposed female mouse lung cells also had much higher PD-L1 expression levels than the PBS-treated control female mouse lung cells although the difference was not statistically significant (p = 0.07) due to the small sample size (Fig. 8A).

Fig. 8.

Infiltrated T cell populations are reduced in chronic chromate exposed-mouse lung tissues. (A) Flow cytometry analysis of PD-L1 protein levels in single cells obtained from lungs of mice exposed to PBS or calcium chromate (100 μg once per week) via oropharyngeal aspiration for 16 weeks. *p < 0.05, compared with the PBS control group (Mean ± SD, n = 4). (B) Flow cytometry analysis of T cell populations in single cells obtained from lungs of mice exposed to PBS or calcium chromate (100 μg once per week) via oropharyngeal aspiration for 16 weeks. *p < 0.05, **p < 0.01, compared with the PBS control group (Mean ± SD, n = 5).

The signal transduction resulting from PD-L1 interaction with PD-1 on T lymphocytes inhibits cytotoxic T cell activation and causes T cell exhaustion in tumor tissues, leading to tumor cell immune tolerance and immune escape, facilitating cancer development and progression [36–38]. Since chronic Cr(VI) exposure increased PD-L1 protein levels in mouse lungs, we further determined whether chronic Cr(VI) exposure had an effect on T cell infiltration into mouse lung tissues. Flow cytometry analysis revealed that the number of T cells in chronic Cr (VI)-exposed male and female mouse lungs were significantly less than the PBS-treated male and female mouse lungs (Fig. 8B). These results suggested that chronic Cr(VI) exposure increased PD-L1 levels and reduced T cell presences in mouse lungs, which may cause tumor cell immune escape to promote lung carcinogenesis.

4. Discussion

Although Cr(VI) carcinogenesis is under active research, the underlying mechanism of Cr(VI) carcinogenesis still has not been well understood. Chronic and persistent inflammation is a significant cancer risk factor [39,40]. While chronic Cr(VI) exposure causes long-lasting inflammation in experimental animal and human lung tissues, how chronic Cr(VI) exposure triggers inflammation and its role in chronic carcinogenesis has not been well understood [8]. In this study, we determined that chronic Cr(VI) exposure activated canonical and non-canonical NFκB pathways to promote inflammatory responses. The lncRNA ABHD11-AS1-mediated non-canonical NFκB pathway activation by chronic Cr(VI) exposure up-regulated immune checkpoint molecule PD-L1 protein levels and reduced T lymphocytes infiltration to lung tissues, resulting in tumor cell immune escape to promote Cr (VI)-induced lung carcinogenesis (Fig. 9).

Fig. 9.

A schematic summary of chronic Cr(VI) exposure activating the non-canonical NF-κB pathway to promote immune checkpoint protein PD-L1 expression and lung carcinogenesis.

A significant hurdle in Cr(VI) carcinogenesis study is that a long time of Cr(VI) exposure was reported to produce lung tumors in experimental animals. To facilitate the mechanism study of Cr(VI) carcinogenesis, we attempted to produce a new mouse model demonstrating the lung carcinogenic effect of Cr(VI) with a shorter term of Cr(VI) exposure. We demonstrated that a 16-week calcium chromate exposure via oropharyngeal aspiration once a week was capable of inducing lung adenocarcinomas in A/J mice. Previous studies reported that 9 months to 2 years were used to observe lung adenocarcinomas in calcium chromate-exposed mice or rats [23,26,27]. The successful establishment of a mouse lung carcinogenesis model with a 16-week calcium chromate exposure represents an important progress in Cr(VI) carcinogenesis research and is expected to accelerate Cr(VI) carcinogenesis mechanistic studies.

However, previous studies showed that chronic chromate exposure mainly induced lung squamous cell carcinoma in humans (chromate workers) and rats [26,41–43]. We think that several factors could contribute to the differences of Cr(VI)-induced lung cancer types among humans, rats and mice: (1) Genetic factors: It is known that it is difficult to establish lung squamous cell carcinoma models in mice. Most genetic modifications or chemical carcinogen exposures using human exposure relevant routes only produced lung adenocarcinomas in mice. (2) Cigarette smoking as a potential confounding factor in chromate workers: Most chromate workers were heavy cigarette smokers [41–43]. It is known that cigarette smoking is a significant contributor to human lung squamous cell carcinoma. (3) Cr(VI) exposure time: Most chromate workers were exposed to chromate for more than 20 years [41–43]. Interestingly, Levy et al. reported that lung tumors in rats exposed to chromate (strontium chromate) for two years via intrabronchial pellet implantation were mostly squamous cell carcinomas [26]. However, Takahashi et al. found that lung tumors in rats exposed to strontium chromate for 9 months via the same intrabronchial pellet implantation procedure were mostly adenocarcinomas [27]. In this study, mice were exposed to calcium chromate via oropharyngeal aspiration once per week for 16 weeks and only lung adenocarcinomas were observed in chromate-exposed mice. This is consistent with previous studies reporting that only lung adenomas and adenocarcinomas were observed in chromate-exposed mice [23,44].

Previous studies have shown that Cr(VI) exposure was capable of activating canonical NFκB pathway, although the underlying mechanism has not been well understood [8,45]. However, the effect of Cr(VI) exposure on the non-canonical NFκB pathway is not known. The findings from this study showed, for the first time, that the non-canonical NFκB pathway was activated in chronic Cr(VI) exposure-transformed human bronchial epithelial cells, and in chronic Cr(VI) exposure-caused mouse and human lung tumor tissues. Our mechanistic studies revealed that chronic Cr(VI) exposure activated the non-canonical NFκB pathway through lncRNA ABHD11-AS1-mediated down-regulation of TRAF3 protein, a key component in the inhibitory complex of the non-canonical NFκB pathway.

How does ABHD11-AS1 down-regulate TRAF3 protein levels? TRAF3 is an adaptor signaling molecule and its protein level is mainly regulated by its posttranslational modification of ubiquitination [46]. The K48-linked polyubiquitination of TRAF3 mediated by cIAP1/2 or other E3 ligases promotes TRAF3 proteasome degradation, which is important for non-canonical NFκB pathway activation. On the contrary, deubiquitylation or removal of K48-linked polyubiquitin from TRAF3 by deubiquitinases blocks TRAF3 degradation and impairs non-canonical NFκB pathway activation. In this study, we identified USP15 as a new deubiquitinase that removed K48-linked polyubiquitin from TRAF3 and reduced its proteasome degradation. This was evidenced by: (i) our co-immunoprecipitation experiments showed that USP15 directly interacted with TRAF3; (ii) stably expressing the antisense RNA of ABHD11-AS1 reduced TRAF3 protein ubiquitination levels, however, knocking down USP15 greatly increased TRAF3 protein ubiquitination levels, leading to TRAF3 protein level down-regulation and the non-canonical NFκB pathway activation. Our recent study showed that chronic Cr(VI) exposure up-regulated the expression of ABHD11-AS1, which interacted with SART3 to promote USP15 nuclear localization and regulate RNA alternative splicing [22]. It is likely that ABHD11-AS1-promoted USP15 nuclear localization reduced cytoplasmic USP15 availability and activity towards TRAF3 leading to increased TRAF3 proteasome degradation and non-canonical NFκB pathway activation.

The canonical NFκB pathway signaling is known as the master regulator of inflammation and immunity, playing critical roles in cancers [47,48]. Accumulating evidence indicates that non-canonical NFκB pathway activation also plays important roles in inflammation, immunity, and carcinogenesis [16,17]. This study showed that chronic Cr(VI) exposure was capable of activating both canonical and non-canonical NFκB pathways. Previous studies reported that interactions existed between canonical and non-canonical NFκB pathways, which facilitated inflammatory cytokine expressions to promote inflammation [49,50]. It is thus likely that both canonical and non-canonical NFκB pathways may play important roles in mediating chronic Cr(VI) exposure-caused chronic inflammation and promoting lung carcinogenesis.

Histology analyses revealed that a 16-week-Cr(VI) exposure caused acute and chronic inflammation in mouse lungs in proximity to areas of adenocarcinomas with significant macrophage infiltration. One of the mechanisms of chronic Cr(VI) exposure causing inflammation in this study could be the up-regulation of inflammatory cytokine IL-6 expression in chronic Cr(VI)-exposed human bronchial epithelial cells and mouse lung tissues. IL-6 signaling activates the Jak/Stat3 pathway to promote cell survival, proliferation, malignant transformation and inflammation [51,52]. Indeed, a high activation of the Jak/Stat3 pathway in chronic Cr(VI) exposure-transformed cells and Cr(VI) exposure-induced mouse and human lung tumors tissues was observed as evidenced by their significantly higher levels of phospho-Stat3 than the corresponding control cells and normal lung tissues. While IL-6 expression was known to be regulated by the canonical NFκB pathway activation, the findings from this study showed that the lncRNA ABHD11-AS1/non-canonical NFκB pathway signaling also contributed significantly to the upregulation of IL-6 level, representing a new mechanism that regulates IL-6 expression.

In addition to promoting inflammation, non-canonical NFκB pathway signaling also regulates immunity [17]. A recent study showed that RelB was capable of binding to the promoter region of CD274 gene to increase CD274 transcription and up-regulate PD-L1 expression in prostate cancer cells [53]. In this study, we found that chronic Cr(VI) exposure up-regulated PD-L1 expression via RelB acting at the posttranslational level. This was supported by the observations showing (i) Cr(VI) exposure increased PD-L1 protein level but did not affect CD274 mRNA level; (ii) knocking down RelB significantly reduced PD-L1 protein level but did not have a significant effect on CD274 mRNA level; and (iii) inactivation of the non-canonical NFκB pathway by stably expressing the antisense RNA of ABHD11-AS1 in Cr(VI)-transformed cells significantly increased PD-L1 degradation. Moreover, it was further determined that ABHD11-AS1/non-canonical NFκB pathway up-regulated PD-L1 expression level by the IL-6/Jak signaling to increase PD-L1 protein stability. These findings revealed a new mechanism by which non-canonical NFκB pathway activation regulates PD-L1 expression.

PD-L1 is an immune checkpoint molecule. It is now well accepted that PD-L1 interaction with PD-1 on T cells causes T cell exhaustion in tumor microenvironment leading to tumor cell immune evasion and promoting cancer development and progression. Consistent with the findings showing that PD-L1 protein levels are significantly higher in Cr (VI)-transformed cells and chronic Cr(VI) exposure-induced mouse and human lung tumor tissues, chronic Cr(VI)-exposed mouse lung tissues had significantly less T cells than the PBS-exposed control mouse lungs. These findings provided evidence supporting an important role of PD-L1 up-regulation-mediated immunosuppression in Cr(VI) carcinogenesis.

In summary, this study established a new animal model demonstrating the lung carcinogenic effect of a 16-week chromate exposure via oropharyngeal aspiration. This study identified USP15 as a new deubiquitinase that interacted with TRAF3 and regulated TRFA3 protein levels. Further mechanistic studies revealed that chronic Cr(VI) exposure activated the non-canonical NFκB pathway through the lncRNA ABHD11-AS1/USP15-mediated TRAF3 down-regulation. The non-canonical NFκB pathway activation enhanced the IL-6/Jak/Stat3 signaling to induce inflammation and stabilize PD-L1 protein, and reduced T lymphocyte infiltration to the lungs, promoting chronic Cr (VI) exposure-caused lung carcinogenesis.