Background
Lung cancer morbidity and mortality continue to exhibit progressive increases year by year, with studies highlighting it as the most common malignant tumor worldwide of which lung adenocarcinoma (LAD) represents the most common histological subtype of lung cancer [
1]. Tobacco consumption is the main cause of lung cancer, and in addition, genetic factors, diet, occupational exposure, as well as air pollution have all been implicated in the epidemiology of lung cancer independently or in concert with smoking [
2]. Although some progress has been made in the arena of lung cancer therapeutics, the overall survival is still relatively low owing to earlier chemotherapy resistance and late stage diagnosis [
3]. A greater understanding of the finer molecular mechanism underlying LAD cell proliferation and metastasis is crucial in order to prevent tumor metastasis and improve patient survival.
As a member of the HOX family, Homeobox A10 (HOXA10) has been shown to play a notable role in the regulation of various cellular functions in various kinds of diseases and cancers, including gastric cancer, nasopharyngeal carcinoma, and prostate carcinoma [
4‐
6]. Moreover, HOXA10 has been demonstrated to promote the occurrence of LAD [
7]. Histone deacetylase 1 (HDAC1) represents a deacetylase that has been implicated in the occurrence and development of various cancer in addition to exerting potential functions on cell functions [
8]. A relationship between HDAC1 and lung cancer has been previously emphasized, with HDAC1 suggested as a functional diagnostic and prognostic indicator of lung cancer [
9]. Functionally speaking, HOXA10 has been identified as a stimulatory factor for the transcription of HDAC1 [
10]. Du Z et al., demonstrated that HDAC1 can stabilize DNA methyltransferase 1 (DNMT1), the primary enzyme maintaining DNA methylation [
11]. Moreover, via methylation means, DNMT1 has been shown to inhibit the expression of Krüppel-like factor 4 (KLF4) [
12], a member of KLF transcription factor family, which inhibited the invasion and metastasis of LAD cells [
13]. Based on the aforementioned exploration of evidence, we put forward the hypothesis that HOXA10 promotes the development of LAD through downregulating KLF4 mediated by histone deacetylase HDAC1 and DNMT1. In order to prove our hypothesis, in vitro assays as well as in vivo experiments were performed to elucidate the interactions among HOXA10, HDAC1, DNMT1, and KLF4 in LAD in an attempt to identify the detailed potential mechanism underlying the treatment of LAD.
Methods
Clinical sample collection
From October 2017 to December 2018, 42 cases of LAD and paracancerous tissues (> 3 cm away from tumor tissues) were collected from patients diagnosed with LAD (25 males and 17 females), all of whom were pathologically confirmed and surgically treated at the Second Hospital of Jilin University. Of the 42 cases, 29 cases were under 55 years and 13 cases were over 55 years. The tumor tissues were either classed as well differentiated (
n = 20), moderately differentiated (
n = 12), or poorly differentiated (
n = 10). Moreover, lymph node metastasis was detected in 16 cases while no such findings were identified in the remaining 26 cases. According to latest staging system for lung cancer published by the International Union Against Cancer and American Joint Committee on Cancer [
14], the enrolled cases were further divided into stage II (
n = 27) and stage IIIa (
n = 15). Patients who received any medication, radiochemotherapy or immunobiological therapy were excluded from the study.
Cell culture
LAD cell lines GLC-82 (ZY-H065), XWLC05 (HTX-2446), and SPCA-1 (BSC-5307481030-01) were purchased from the American Type Culture Collection (Manassas, VA, USA), H1299 (3111C0001CCC000469) and A549 (3111C0001CCC000002) with the human bronchial epithelial (HBE) cells (3111C0001CCC000174) purchased from the Cell Resource Center of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China). GLC-82, SPCA-1, and HBE cells were cultured in Dulbecco’s modified Eagle’s medium (GIBCO-BRL, CA, USA) containing 10% fetal bovine serum (FBS) (GIBCO-BRL), H1299 and XWLC05 cells were cultured in Roswell Park Memorial Institute Medium 1640 medium (Shanghai Yuanmu Biotechnology Co., Ltd., Shanghai, China) containing 10% FBS, and A549 cells were cultured in McCoy’s 5A Media (Modified with Tricine) containing 10% FBS.
Cell transfection
Three short hairpin RNAs (shRNAs) targeting HOXA10, HDAC1, DNMT1, and KLF4 in addition to one scramble shRNA against negative control (sh-NC) were cloned into pLKO.1-puro (Addgene, Cambridge, MA, USA), respectively. The shRNAs and lentiviral packaging vectors were co-transfected into HEK293T cells using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA) as per the manufacturer’s protocol for 48–72 h, after which the supernatant was collected to infect target cells at later stage. The coding sequences of HDAC1 and DNMT1 were amplified, digested with
HindIII/
EcoRI (Thermo Fisher Scientific, Waltham, MA, USA), and inserted into
HindIII/
EcoRI-digested pcDNA3.1 (+) (Addgene). HOXA10-pcDNA3.1 was transfected into the LAD cells with the overexpression effect subsequently confirmed by Western blot analysis 48 h post transfection. The shRNA sequences are depicted in Table
1.
Table 1
Interference sequences for cell transfection
HOXA10 | shRNA-1: 5′-CTTTCGCGCAGAACATCAA-3′ |
shRNA-2: 5′-TATGTACCTTACTCGAGAG-3′ |
shRNA-3: 5′-TGAATCGAGAAAACCGGAT-3′ |
HDAC1 | shRNA-1: 5′-AGCGACGACTACATCAAATTC-3′ |
shRNA-2: 5′-ATGGCTATACCATCCATAATG-3′ |
shRNA-3: 5′-AGACCCTGACAAACCAATTTC-3′ |
DNMT1 | shRNA-1: 5′-CCAUGAGCACCGUUCUCCTT-3′ |
shRNA-2: 5′-GGAGAACGGUGCUCAUGGTT-3′ |
shRNA-3: 5′-TTGATGTCAGTCTCATTGG-3′ |
KLF4 | shRNA-1: 5′-TACCCATCCTTCCTGCCCGAT-3′ |
shRNA-2: 5′-ATCGGTCATCAGCGTCAGCAA-3′ |
shRNA-3: 5′-AAGTCATCTTGTGAGTGGATAA-3′ |
scrambled shRNA | 5′-GCGATGGGCGAACTGACACG-3′ |
Reverse transcription quantitative polymerase chain reaction (RT-qPCR)
Total RNA was extracted from the tissues using TRIzol (Invitrogen, Carlsbad, CA, USA), after which cDNA was obtained through reverse transcription using a reverse transcription kit (RR047A, Takara, Japan). RT-qPCR was performed using a SYBR® Premix Ex TaqTM II (Perfect Real Time) kit (DRR081, Takara, Japan) in a real-time PCR instrument ABI 7500 (ABI, Foster City, CA, USA). Primers were synthesized by Shanghai Biotech Co., Ltd. (Shanghai, China) (Table
2). The Ct value of each well was recorded, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) employed as the internal reference. The 2
-ΔΔCt formula was applied to calculate relative expression.
Table 2
Primer sequences for RT-qPCR
HOXA10 | F: 5′-AGAGATTAGCCGCAGCGTCC-3′ |
R: 5′-TTCCTGGGCAGAGCCTGAAG-3’ |
HDAC1 | F: 5′-CTACTACGACGGGGATGTTGG-3′ |
R: 5′-GAGTCATGCGGATTCGGTGAG-3′ |
DNMT1 | F: 5′-AACCTTCACCTAGCCCCAG-3′ |
R: 5′-TGACAGGTGGTCACTCCTCATG-3′ |
KLF4 | F: 5′-TTCTCCACGTTCGCGTCCGG-3′ |
R: 5′-TCTCGCCAACGGTTAGTCGGGG-3′ |
GAPDH | F: 5′-CACCCACTCCTCCACCTTTG-3′ |
R: 5′-CCACCACCCTGTTGCTGTAG-3′ |
Western blot analysis
The LAD cells and tissues were lysed using radioimmunoprecipitation assay lysis buffer (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China), after which the protein concentration was measured using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific). The protein was subsequently separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to a nitrocellulose membrane (Millipore, Bedford, MA, USA) which was blocked, and incubated with the primary antibodies overnight. Protein bands were detected using an enhanced chemiluminescence detection kit (Pierce Biotechnology, Rockford, IL, USA). The used primary antibodies were anti-HOXA10 (ab23392, 1: 500), anti-HDAC1 (ab7028, 1: 2000), anti-DNMT1 (ab134148, 1: 1000), anti-KLF4 (ab215036, 1: 1000), anti-HA (ab130275, 1: 1000), anti-GAPDH (ab181602, 1: 10000), and the secondary antibody was IgG (ab6721, 1: 2000). All mentioned antibodies were obtained from Abcam (Cambridge, MA, USA). Image J software (
http://rsb.info.nih.gov/ij/, Bethesda, MD, USA) was used to quantify the gray scale of the detected protein.
Cell counting kit-8 (CCK-8) assay
The transfected LAD cells were seeded into 96-well plates at a density of 3 × 103 cells/well, then, after 0, 24, 48 and 72 h, CCK-8 solution (Signalway Antibody, College Park, MD, USA) was added to each well and incubated for 1 h. The absorbance (optical density value) was measured at a wavelength of 450 nm on a microplate spectrophotometer (Thermo Fisher Scientific).
Flow cytometry analysis of cell cycle and apoptosis
Cell cycle analysis was performed using cells that were treated with antibody to 5-bromo-2′-deoxyuridine-fluorescein isothiocyanate (FITC) (BD Biosciences, Franklin Lakes, NJ, USA) and DNA was stained with 7-AminoactinomycinD (7-ADD, Sigma-Aldrich Chemical Company, St Louis, MO, USA) [
15].
Cell apoptosis analysis was performed in line with an Annexin V-FITC kit (Shanghai Beyotime Biotechnology Co., Ltd., Shanghai, China) which was used for staining, with both the DNA content and apoptosis analyzed using a FACScan flow cytometer (BD Biosciences).
Scratch test
When the cell confluence reached approximately 90%, sterile pipette tips were used to scratch the middle of the cells. The cells were cultured in serum-free medium for 24 h and evaluated accordingly. ImagePro Plus analysis software 7.0 (Media Cybernetics, Inc., Rockville, MD, USA) was employed to determine the migration distance.
Transwell assay
The upper chamber of the Transwell chamber was covered with Matrigel (50 mL; 356,234; Becton, Dickinson and Company, NJ, USA). The transfected cells were then seeded into the upper cavity of the Transwell filter membrane and incubated for 48 h in serum-free medium. The infiltrated cells on the underside of the filter membrane were fixed with 5% glutaraldehyde, stained with 0.1% crystal violet and counted under the microscope. The invasion was determined by the number of cells that passed through the Matrigel.
Dual luciferase reporter gene assay
The binding site on the HDAC1 promoter to HOXA10 and the binding site on the KLF4 promoter to DNMT1 were mutated by PCR. The wild-type (WT) and mutant (MUT) sequences of the HDAC1 or KLF4 promoter were inserted into the pGL3-vector (Promega Corporation, Madison, WI, USA). The luciferase vector pRL-TK (Promega) and pGL3-HDAC1 promoter vector were transfected into the cells in the presence of sh-HOXA10 or sh-NC, respectively. Similarly, the pRL-TK and pGL3-KLF4 promoter vectors were transfected into the cells in the presence of sh-DNMT1 or sh-NC. After 48 h had elapsed, the luciferase activity was determined using a Dual Luciferase Reporter Assay System (Promega) normalized to Renilla luciferase activity. The activity control for co-transfection with the pRL-RK and pGL3-HDAC1 promoters was set as 1.0.
Chromatin immunoprecipitation (ChIP)
A ChIP kit (Millipore) was used to investigate the enrichment of HOXA10 in the HDAC1 promoter region and the enrichment of DNMT1 in the KLF4 promoter region. Cells at the logarithmic growth phase from group were collected, fixed with formaldehyde and sonicated. The negative control IgG antibodies (ab205718, 1: 50, Abcam) and target protein-specific antibody HOXA10 (ab23392, 1: 500, Abcam) respectively. DNMT1 (ab13537, 1: 50, Abcam) were added for incubation at 4 °C overnight. After that, endogenous DNA-protein complexes were precipitated and de-crosslinked. Finally, DNA fragments were extracted and purified, and used for testing the binding of HOXA10 to HDAC1 promoter and the binding of DNMT1 to KLF4 promoter. The primer sequences are shown in Table
3.
Table 3
Primer sequences for ChIP
HDAC1 | F: 5′-AAAGAAAGGAAACCTGCCCTC-3′ |
| R: 5′-TGCAGTCACCCAGGATGACTA-3′ |
KLF4 | F: 5′-CCTGACCATGAAAACTGTGAGATA-3′ |
| R: 5′-GCTGGTCTTGAACTCCTGCGCTCA-3′ |
Co-immunoprecipitation (co-IP)
Co-IP analysis was performed using Pierce™ Co-IP kit (Thermo Scientific Pierce, 26,149). First, rabbit anti-human HDAC1 antibody (ab150399, 1: 100, Abcam) and goat anti-rabbit IgG (ab136636, 1: 5000, Abcam) were conjugated with AminoLink Plus Coupling Resin and incubated at room temperature for 120 min. The A549 cells were then lysed using IP lysis buffer solution, with the cell lysates harvested, pre-treated with Pierce Control Agarose Resin and added into the antibody-crosslinked resin. After co-IP at 4 °C overnight, the precipitate was eluted and subjected to Western blot analysis.
In vitro acetylation assay
Glutathione S-transferase-fused DNMT1 was co-transfected into A549 cells with either sh-NC or sh-HDAC1, and 50 μL acetyltransferase assay buffer was incubated in 20 μM acetyl-CoA at 30 °C for 2 h. The reaction mixture was analyzed by SDS-PAGE and further examined through Western blot analysis.
In vivo ubiquitination (Ub) test
The A549 cells in the presence of sh-NC and sh-HDAC1 were co-transfected with HA-Ub and pCMV5-myc-DNMT1 plasmids for 2 h, and treated with 10 μg/mL MG132 (Sigma-Aldrich). The cell lysate was then immunoprecipitated with anti-myc antibody (ab172, 1: 250, Abcam), with the protein mixture in the precipitate was separated by SDS-PAGE. Ub-DNMT1 was analyzed with antibody to HA (ab130275, 1: 150, Abcam). The cell lysate was then immunoprecipitated with antibody to DNMT1, and the ubiquitination level of DNMT1 protein in each experimental group was detected by Western blot analysis of antibody to HA.
Pulse-chase
The supernatant was collected through the pre-spin using 40 μL Agarose A/G beads (Upstate Biotechnology, NY, USA), added with the primary antibody, and immunoprecipitated at 4 °C overnight. Protein samples were denatured in Laemmli buffer (4 ×), separated on a 10% polyacrylamide gel and subjected to Western blot analysis as described above.
Methylation-specific PCR (MSP)
MSP was performed on sodium bisulfate-treated DNA using the EZ DNA Methylated Gold™ Kit (Zymo Research, Irvine, CA, USA) based on the manufacturer’s instructions. As previously mentioned, methylation-specific primers for the KLF4 promoter were designed using the MethPrimer program. The methylation-specific primers and SYBR-Green reaction mixture were applied for PCR amplification of bisulfite-converted genomic DNA. The primers used are illustrated in Table
4.
Table 4
Primer sequences for methylation-specific PCR
KLF4 (methylated) | F: 5′-CGTAGGGTTTAAATAGGTGATAACG-3′ |
R: 5′-AAATAATAAAAACTCGAACACCGAA-3′ |
KLF4 (nonmethylated) | F: 5′-TGTAGGGTTTAAATAGGTGATAATGA-3′ |
R: 5′-AAATAATAAAAACTCAAACACCAAA-3′ |
Xenograft in nude mice
Eighteen six-week-old female athymic BALB/c nude mice (weight: 18–20 g) were purchased from the Experimental Animal Center of Guangdong Province (Guangdong, China), and reared under specific pathogen-free conditions. Nude mice were intraperitoneally injected with LAD cells (1 × 106 cells/200 μL/mouse) transfected with sh-NC and oe-NC, with sh-HOXA10 and oe-NC, in addition to sh-HOXA10 and oe-HDAC1. The longest diameter (L) and shortest diameter (W) of the tumors in the mice were measured every 4 days post injection with the tumor growth plotted and a data curve constructed based on the formula V = L × W2 × 0.5. On day 25, the mice were euthanized and their tumors were excised. Six tumor samples from each group were weighed and averaged.
Immunohistochemistry
LAD tissue sections were dewaxed with xylene and hydrated with alcohol at descending concentrations. Following antigen retrieval, the tissues were incubated with goat polyclonal antibody to HOXA10 (1: 200; ab191470; Abcam) at 4 °C overnight, as well as with biotinylated rabbit anti-goat secondary antibody (ab97100; Abcam) at 37 °C for 30 min. Diaminobenzidine (Bost Biotechnology Co., Ltd., Wuhan, China) was added stained for 1–2 min, while hematoxylin (KeyGEN Biotech Co., Ltd., Nanjing, Jiangsu, China) was counterstained for 1 min. The tissue sections were subsequently dehydrated and fixed using neutral balm. Five fields of view were selected and observed under an optical microscope (200 ×, Nikon, Tokyo, Japan) with 100 cells from each field analyzed. Brown-yellow cells were considered to be positive. Cells with a staining degree greater than 25% were considered to be positive cells with the positive rate calculated using the following formula: positive rate = (the number of positive cells/the number of total cells) × 100%.
Statistical analysis
SPSS 22.0 statistical software (IBM Corp. Armonk, NY, USA) was applied for statistical analysis. All data were expressed as the mean ± standard deviation. Data between LAD and paracancerous tissues were compared by paired t-test, while data between the other two groups of unpaired design were analyzed using unpaired t-test. Data comparison among multiple groups was conducted using one-way analysis of variance (ANOVA) and Tukey’s post-test. Cell viability at different time points was compared using two-factor ANOVA, and tumor volume data at different time points were compared using Bonferroni-corrected repeated measures ANOVA. Pearson correlation or Spearman correlation was employed to analyze the correlation between HOXA10 and HDAC1. p < 0.05 was considered to be indicative of statistically significant difference.
Discussion
LAD represents the most common lung cancer subtype, and is often accompanied with a high rate of morbidity and mortality [
18]. In spite of commendable advances in LAD therapy, patients are often diagnosed with this disease at an advanced or metastatic stage, resulting in a 5-year survival of less than 20% [
19]. Thus, it is absolutely necessary to identify the genes and mechanisms related to LAD in order to improve diagnosis and treatment efficiency. In recent years, several articles have reported the regulation of HOXA10 in LAD [
16,
20,
21]. Hence, the current study aimed to elucidate the downstream mechanism by which the HOXA10 influences and contributes to LAD. Our results demonstrated that HOXA10 promotes the malignant phenotypes of LAD via regulation of the histone deacetylase HDAC1-mediated DNMT1/KLF4 axis.
HOXA10 plays an important role in regulating cell differentiation, maturation, development and proliferation, and has been implicated in the occurrence and development of certain cancers [
22‐
24]. A key initial finding of our study detected high levels of HOXA10 expression in LAD tissues and cell lines. We subsequently set out to investigate its effect on LAD phenotypes, and the results indicated that following HOXA10 silencing, cell viability, migration, and invasion were inhibited, cell cycle was arrested, and apoptosis increased. This finding was consistent with the report that HOXA10 overexpression may play an essential role in non-small cell lung carcinoma (NSCLC) tumorigenesis [
20]. A previous study demonstrated that ELK1 induced upregulation of long noncoding RNA HOXA10-AS, which in turn activated the Wnt/β-catenin signaling and promoted LAD progression [
21].
Next, the downstream mechanism of HOXA10 regulating lung cancer proliferation, migration, and invasion was further investigated. Our results illustrated that HOXA10 could directly bind to HDAC1 and subsequently promote its expression in LAD. Minamiya Y et al., emphasized higher HDAC1 expression as an independent marker of poor prognosis in patients with LAD [
25]. In addition, HOXA10 silencing suppresses the proliferation of hepatoma cells, induces cell cycle arrest and apoptosis by upregulating HDAC1 transcription [
10], which was consistent with our findings.
Johann C Brandes et al. concluded that class I HDACs are mediators that stabilize DNMT1 and are promising targets for the prevention of lung cancer induced by smoke carcinogens [
26]. Subsequent results in our study revealed that the expression of DNMT1 was higher in tumor samples, highlighting the correlation between DNMT1 and LAD. Consistent with our observations, a previous study indicated that DNMT1 was overexpressed in primary NSCLCs [
27]. We then speculated that HDAC1 could potentially promote the proliferation, migration and invasion of LAD via DNMT1, which we subsequently verified through a HDAC1 silencing experiment as well as a DNMT1 rescue experiment. We then explored the interaction between HDAC1 and DNMT1 and our data indicated that HDAC1 inhibited the degradation of DNMT1 by strengthening its deacetylation, a finding of which was consistent with existing literature that suggests that DNMT1 was stabilized by HDAC1 [
11].
Bearing in mind the function of DNMT1 as a DNA methyltransferase, we subsequently set out to identify the catalytic substrate of DNMT1. Previous research has suggested that upregulation of DNMT1 and promoter hypermethylation may triggered a downregulation in the expression of KLF4 [
28]. Additionally, KLF4 is found to act as important player in the progression of many aggressive cancers, such as lung cancer [
29,
30]. Our results provided further evidence attesting the notion that KLF4 was indeed associated with LAD, with poorly expressed KLF4 identified in LAD. Functionally, we also observed that by silencing DNMT1, KLF4 methylation was increased as reflected by activation of its expression. Moreover, downregulation of DNMT1 could suppress cancer cell viability, migration and invasion through upregulation of KLF4 expression. Our observation was similar to the description that DNMT1 might reduce KLF4 expression by catalyzing DNA methylation in the promoter region of KLF4 [
31]. Furthermore, DNMT1 inhibition reduces KLF4 promoter DNA methylation and activates KLF4 expression in pancreatic cancer cells [
28].
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.