Abnormal expression and potential clinical value of oncogenic Krüppel-like factor-5 in lung squamous cell carcinoma

Abstract

BACKGROUND

Krüppel-like factor-5 (KLF5) is a zinc-finger transcription factor related to tumor progression. However, the relationship between KLF5 and lung cancer remains to be identified.

AIM

To investigate the clinical value of KLF5 and interference with KLF5 mRNA transcription on the effects of biological behaviors in lung squamous-cell carcinoma (LUSC).

METHODS

Lung KLF5 mRNA data were extracted from bioinformatics databases. Blood and tissues from a cohort of patients with benign or malignant lung diseases were collected with ethical committee consent to validate KLF5 expression via multiplex immunofluorescence and immunohistochemistry, Western blot, Enzyme-Linked Immunosorbent Assay or quantitative polymerase chain reaction. Furthermore, KLF5 mRNA was silenced in lung A549 cells to validate biological behaviors in vitro and nude mouse xenograft growth in vivo, respectively.

RESULTS

A cohort of bioinformatics databases revealed high KLF5 mRNA expression in LUSC (P < 0.001) but lower KLF5 mRNA expression in lung adenocarcinoma. Upregulated KLF5 in the lung or sera of patients with lung cancer (P < 0.001) were confirmed that related to poor differentiation, lymph node or distant metastasis. Furthermore, the incidence of KLF5 levels greater than 500 ng/mL in LUSC patients was 86.7%, which was significantly greater (P < 0.001) than that in cases with benign lung diseases (13.3%) or healthy controls. Functionally, silencing KLF5 mRNA with a specific shRNA significantly suppressed A549 cell proliferation, decreased cell migration, increased the ratio of G2 phase cells in vitro, and inhibited the growth of nude mouse xenografts in vivo.

CONCLUSION

KLF5 is a novel diagnostic biomarker or potential therapeutic target for LUSC.

Key Words: Targeted therapy; Xenograft growth; Biological behaviors; Diagnostic biomarker; Lung adenocarcinoma; Lung squamous-cell carcinoma; Lung cancer; Oncogenic Krüppel-like factor-5

Core Tip: Krüppel-like factor-5 (KLF5) is upregulated in lung cancer, with stronger expression in lung squamous-cell carcinoma (LUSC) than in lung adenocarcinoma. Upregulated KLF5 expression was associated with the proliferation, migration and cell cycle of lung cancer cells, which is likely a favorable diagnostic biomarker of LUSC and could be a new molecular target for lung cancer. Furthermore, KLF5 might exert its oncogenic regulatory effects on the growth of lung cancer by increasing metastasis or progression by activating related signaling pathways.

INTRODUCTION

Non-small cell lung cancer (NSCLC) remains the leading cause of cancer-related death worldwide[1]. Lung adenocarcinoma (LUAD) or lung squamous-cell carcinoma (LUSC) are the most common subtypes of NSCLC[2]. The early symptoms of lung cancer are not obvious, early diagnosis is difficult, the disease can easily metastasize or produce multidrug resistance, and the prognosis is poor[3]. These characteristics are primarily due to the complex biological features of lung cancer[4]. The formation and progression of lung cancer are related to oncogenic or key molecular activation, tumor suppressor gene inactivation, protein functional effector sncRNAs (pfeRNAs)[5], tissue inflammation and the microenvironment[6,7]. Currently, lung cancer research focuses on discovering tumor markers involved in regulating the cell cycle, apoptosis, angiogenesis, or targeted therapy and immunotherapy[8-10]. Recently, increasing evidence has suggested that some members of the Krüppel-like transcription factor (KLF) subfamily of the zinc finger protein (ZFP) superfamily are closely associated with promoting or inhibiting lung cancer progression and are expected to become new biomarkers or therapeutic targets[11,12].

KLFs, such as oncogenic KLF5, are involved in many biological processes and many diseases, especially cancers[13-15]. The KLF5 gene is located at 13q21; its protein consists of 457 amino acid residues, and the terminus contains a transcript activation or repression domain; it is involved in the regulation of cellular behavior, organ injury repair, and tumor progression[16,17]. Several transgenic models have revealed the physiological and pathological functions of KLF5 in cancers[17,18]. KLF5 is involved in diverse oncogenic signaling pathways and is implicated in regulating the differentiation, proliferation, migration, and apoptosis of cells and in tissue remodeling during cancer progression[19,20]. However, the role of KLF5 in lung cancer remains to be identified[21]. The aims of this study were to investigate KLF5 transcription in lung cancers via bioinformatic databases, verify these findings in clinical specimens and investigate the effects of KLF5 mRNA activation on cell biological behaviors in vitro and the growth of nude mouse xenograft tumors in vivo.

MATERIALS AND METHODS

KLF5 bioinformatic databases

KLF5 mRNA expression data in lung cancer or adjacent tissues were extracted from the Gene Expression Profiling Interactive Analysis (http://GEPIA.cancer-pku.cn, GEPIA)[22] and the Cancer Genome Atlas (https://www.cancer.gov, TCGA)[23], which provide crucial interactive and customizable functions, such as the differential levels of KLF5 in LUAD or LUSC and their adjacent tissues, patient survival and correlation. Datasets downloaded from GEPIA or TCGA, including LUAD or LUSC with adjacent tissues were used. Kyoto Encyclopedia of Genes and Genomes (https://www.KEGG.jp, KEGG) pathway enrichment analysis[24] revealed that major signaling pathways are involved in the promotion of cancer progression by KLF5. |R| > 0.1 and P < 0.05 were considered to indicate significant differences.

Patient enrollment

For this study, blood and lung cancer tissues and their adjacent tissues (Adj-Ca) were obtained from a cohort of 120 patients with lung cancer who underwent surgery from July 2015 to December 2018. All patients provided signed informed consent, and the study was approved by the Hospital Ethics Committee (No. 2020-L151, The Second Affiliated Hospital of Nantong University, Nantong 226009, China) and performed in accordance with the medical ethics of the Helsinki Declaration. All patients had complete clinical records and follow-up data and had not received radiotherapy or chemotherapy before surgery. Specimens were collected from patients (81 males and 39 females, 60.2 ± 7.5 years old) diagnosed with LUAD (n = 60, 30 males and 30 females) or LUSC (n = 60, 51 males and 9 females). The tumor node metastasis (TNM) stages of the cohort included 81 samples at Stage I, 17 samples at Stage II, and 22 samples at Stages III-IV. Each sample was divided into two parts for hematoxylin & eosin (H&E) staining, immunohistochemistry (IHC) and multiplex immuno-fluorescence (MIF). The diagnostic criteria were based on the guidelines for the standardized diagnosis and treatment of lung cancer (2019)[25] and were verified by pathological examination. In addition, blood was collected from 60 patients with chronic pneumonia (CP) and 60 healthy controls (NC) from the Nantong Central Blood Bank, Jiangsu, China.

Tissue microarray

A lung tissue microarray (TMA) chip with 2.0-mm lung tissue cores was constructed from 120 cancerous and autologous adjacent tissues (SHYJS-CP-1910013, Shanghai Outdo Biotech Co., China). Briefly, original tissues were fixed in buffered formalin, embedded in paraffin and stained with H&E. to confirm areas of the tumor. The samples were then punctured with a machine to prepare a cylindrical TMA with a 1.5 mm pore diameter and 0.2 mm core spacing on polylysine-coated glass slides fixed at 55 °C for 10 minutes. The samples were fixed before being embedded in paraffin at -20 °C for 5 minutes and sectioned on a microtome. The samples were then moved to a H₂O bath at 45 °C for 2 minutes, unrolled in cold H₂O, maintained at 60 °C for 3 minutes and 58 °C for 18 hours, and stored at -20 °C. Sections were continuously cut from the TMA blocks, and the core sectioning was repeated two times to ensure reproducibility. The tumor content was subsequently confirmed by two pathologists.

MIF

MIF staining was performed with an OPAL IHC kit from PerkinElmer (ABSIN, Shanghai, China) and blocked with antibody diluent, followed by incubation with primary anti-KLF5 (1:1000, Abcam, United Kingdom) and anti-Wnt3a (1:1000, Abcam, United Kingdom) antibodies for 1 hour; negative controls were incubated without antibody. Detection was based on an OPAL Polymer HRP antibody (Waltham, MA, United States), and signals were visualized with OPAL tyramide signal amplification plus agent, after which the section was placed in EDTA buffer (pH 8.0) and heated in a microwave. After staining, the samples were washed, and the samples were subsequently sealed with glycerin at 25 °C in the dark. Blinded evaluations were performed simultaneously and independently by two pathologists, and scores were calculated based on the intensities and numbers of positive cells via Image-Pro Plus 6.0 software (Rockville, MD, United States).

IHC

The results of the IHC staining of KLF5 on the TMA after hydration, hydrogen peroxide blocking, high-pressure heat antigen retrieval and animal serum blocking were analyzed via an immunohistochemical kit (Bausch & Lomb, United States). Anti-human KLF5 antibodies (1:1000; Abcam, United Kingdom) were incubated at 4 °C overnight. After the samples were washed with phosphate-buffered saline (PBS), biotin-labeled secondary antibodies were added, the samples were rinsed, streptomycin/peroxidase was added, and the samples were rinsed. The primary or secondary antibodies were replaced with PBS, and the samples were photographed under a light microscope. Five high-power visual fields (400)/slices were selected to count KLF5-positive cells, which were classified as 0%-10%, negative (-); 11%-25%, weak (+); 26%-50%, medium (++); and over 51%, strong (+++) staining. The percentage of positive cells under 25% was low, and more than 26% had stronger KLF5 expression.

Western blotting

Protein was purified from cells or lung tissues using RIPA lysis buffer supplemented with phenylmethylsulfonyl fluoride (PMSF; Biotite Biotechnology, Shanghai, China). Protein concentrations were determined using a bicinchoninic acid kit (BCA, Shanghai Beyotime, China). Subsequently, the samples were lysed on ice in radioimmunoprecipitation assay (RIPA) lysis buffer [1 × PBS, 1% NP-40, 0.1% sodium dodecyl sulfate (SDS), 5 mmol/L EDTA, 0.5% sodium deoxycholate, and 1 mmol/L sodium orthovanadate] containing 100 μg/mL PMSF and protease inhibitors (KeyGen, Nanjing, China). Equivalent amounts of protein (50 μg/Lane) were separated via 10% SDS-polyacrylamide gel electrophoresis at 80 V for 40 minutes, followed by 120 V for 1 hour. The proteins were then transferred to polyvinylidine difluoride membranes (Millipore, MA, United States) at 300 mA for 120 minutes, blocked in 5% BSA in blocking buffer (Solarbio, China), and incubated with primary rabbit anti-human KLF5 (1:1000; Abcam, United Kingdom) or anti-GAPDH antibodies (1:1000; CST, United States) overnight at 4 °C, followed by incubation with secondary horseradish peroxidase-conjugated goat anti-rabbit antibodies (1:1000; Abbkine, China) for 2 hours at 25 °C. Bands were visualized via an enhanced chemiluminescence system (ECL, Shanghai Beyotime, China). Images were taken with Quantity One software (Bio-Rad, Inc., United States).

Cell culture

The cells were cultured at 37 °C in medium supplemented with 10% fetal bovine serum (FBS; AusGeneX, Brisbane, Australia), 1% penicillin-streptomycin and 5% CO₂. The following culture media were used: Lung cancer NCI-H1650, SPC-A-1, A549 (cat. no. SCSP-503), and NCI-H1975 cell lines and normal lung epithelial Beas-2B cell lines (Institute of Biochem. & Cell Biol., Shanghai, China). When the cell density was greater than 80%, the culture medium was removed, and the cells were washed with PBS (0.01 mol/L, pH = 7.5) and digested with 0.25% EDTA-Trypsin. When the cells were separated and retracted, trypsin digestion was terminated by adding a double volume of medium, centrifugation and resuspension were inoculated in new bottles for 5 minutes, and the samples were subsequently washed on ice with TBS at centrifugation for 5 minutes. Then, 0.5 mL of the supernatant and 1.0 mL of extraction reagent were added, and the mixture was shaken. After centrifugation at 4 °C for 10 minutes, the protein was retained, 1 mL of ethanol was added, the mixture was centrifuged at 4 °C for 3 minutes, the isotonic supernatant was removed, the mixture was dried at 25 °C, trimethylolaminomethane hydrochloric acid was added, and the protein was examined and stored at -20 °C until use. All cells were cultured at 37 °C and were tested for mycoplasma contamination by MycAwayTM Plus-Color One-Step Mycoplasma Detection Kit (40612ES25, YEASEA, China). All cell lines were validated by STR profiling and tested negative for mycoplasma.

Transfection of KLF5i plasmids

A549 cells were divided into three groups: The blank control (Blank), negative control (N-KLF5) and KLF5 intervention (KLF5i) groups. The sequence used for targeting KLF5 was 5’-AGAGAATTTAGATGCAT-3’, which was used to construct KLF5i lentiviral vectors. The cells in the KLF5i group were inoculated into a 6-well plate with fusion rate greater than 70%, and then infected with KLF5i plasmids, whereas those in the N-KLF5 group were infected with lentiviral vectors. The cells were labeled with green fluorescent protein and observed with a fluorescence microscope after 12 hours; the infection efficiency was calculated for subsequent experiments.

Quantitative real-time PCR

Total RNA was extracted from cells using the Invitrogen TRI^zol method (Carlsbad, CA, United States) according to the manufacturer’s instructions. The concentrations were detected, and the total RNA was reverse transcribed into KLF5 cDNA via the PrimeScript RT Reagent Kit (TaKaRa, Shiga, Japan). The KLF5 cDNAs were subsequently amplified with a pair of primers, namely, KLF5-F: 5’-ACACCAGACCGCAGCTCCA-3’ (nt 908-926) and KLF5-R: 5’-TCCATTGCTGCTGTCTGATTTGTAG-3’ (nt 1048-1072), using a SYBR Green Premix Ex Taq^TM kit (TaKaRa, Shiga, Japan), with glyceraldehyde 3-phosphate dehydrogenase (GAPDH-F: 5’-AGCCACATCGCTCAGACAC-3’ and GAPDH-R: 5’-GCCCAATACG ACCAAATCC-3’ as internal references. The levels of mRNA were calculated via the 2^−ΔΔCt method [Δ^Ct = Ct_KLF-5 - Ct_GAPDH].

Cell apoptosis and cell cycle analysis

A549 cells in the logarithmic growth phase were digested, 2000 cells/well from each group were seeded in 3 double-well plates at 100 μL/well, starting from the 2^nd day, Celigo was used to measure the plate once a day for 5 days, and the number of green fluorescent cells in the scanning hole plate was accurately calculated by adjusting the input parameters to plot the proliferation curve at 3 days. The cells were detected via transfection, trypsin digestion, centrifugation at 1300 rpm, the addition of 200 μL of PBS to resuspend the cells, and Annexin V-APC staining at 25 °C for 10 minutes. Apoptotic cells were detected via flow cytometry. When the cell coverage rate was greater than 80%, the collected cells were digested with trypsin, resuspended, and centrifuged for 5 minutes at 1300 rpm, and the supernatant was discarded. The cells were subsequently centrifuged at 4 °C with 75% ethanol for 1 hour, after which the cell cycle was detected via recentrifugation and staining.

Cell scratch and transwell tests

A549 cells in the logarithmic phase were digested by trypsin to prepare a cell suspension and inoculated at 600 cells/well (n = 3) before being cultured for 14 days. Cells were then immobilized by incubation with 4% paraformaldehyde for 30 minutes, dried, and stained with Giemsa before being photographed to count the resulting clones. Digested A549 (1 × 10⁶) cells were added to the wells for the scratch test. The next day, vertical line scratches on the back were made as far as possible perpendicular to the line, the cells were removed, serum-free medium was added, the samples were placed in a 37 °C incubator for 48 hours and photographed. Transwell mixture was diluted 50 times with Matrigel Matrix Gel on ice, and the digested cells were added to the cell compartments with serum-free medium in the upper chambers or complete medium containing 20% FBS in the lower chambers. Cells were cultured for 48 hours and stained with crystal violet. Cotton swabs were removed from the upper chamber without passing through the cells, and the cells were fixed with 4% paraformaldehyde and stained with 1% crystal violet before being photographed in a random field of view.

Xenograft tumor growth

This study was approved by the Animal Guidelines (S20220221-009) of the Animal Care and Use Committee of Nantong University, China. Six-week-old BALB/c nude mice (n = 18) were randomly assigned to the KLF5i, N-KLF5 and blank groups. Briefly, the corresponding A549 cells (2 × 10⁷) were suspended in 200 μL of DMEM and injected subcutaneously into the right scapular area of BALB/c nude mice. Measurement of tumor size through randomization and blinding were applied with calipers every 4 days, and the volume (mm³) was calculated as 0.5 (length × width²). The mice were sacrificed with anesthetized by inhalation of anesthetic ether at the end of the 5^th week. Xenograft tumors were dissected, fixed with 4% paraformaldehyde, embedded in paraffin, deparaffinized in xylene, and dehydrated in gradient ethanol. Finally, pathological examination and IHC and MIF analyses were performed after H&E staining.

Enzyme-linked immunosorbent assay

Serum levels of KLF5 (Qiming Biotechnol Co., Shanghai, China) were independently and quantitatively detected according to the manufacturer’s instructions by two researchers. Absorbance (A) values were measured for repeat readings. Level was calculated as the mean A of the corresponding standard curve. Specific concentrations (pg/mg) of KLF5 were calculated according to the KLF5 Level per milligram of wet lung.

Statistical analysis

Statistical Product and IBM SPSS statistics 23 (IBM Corp., NY, United States) were used for analysis. The data are expressed as the mean ± SD. Differences were compared by one-way ANOVA, χ² tests, and t tests for the groups. KLF5 in tissues and the total survival time of patients were analyzed via the Mann-Whitney U test and Kaplan-Meier analysis. A log-rank test multivariate Cox regression model was used to evaluate the mortality risk ratio and 95% confidence interval (95%CI). When P < 0.05, the difference was considered statistically significant.

RESULTS

Bioinformatics and molecular functions of KLF5

A comparative analysis of KLF5 mRNA in lung cancer (LUSC or LUAD) and normal tissues based on the TCGA and GEPIA databases is shown in Figure 1. Increased levels of KLF5 mRNA were detected in LUAD or LUSC tissues, with abnormally upregulated KLF5 mRNA in LUSC (Figure 1A, P < 0.001). The KLF5 mRNA levels differed between LUSC or LUAD tissues and normal tissues. Specifically, KLF5 mRNA expression was significantly higher in LUSC than in normal tissues (Figure 1B, P < 0.001), but this difference was not significant in LUAD tissues (Figure 1C). Bioprocess-enriched KLF5 mRNAs, covalent chromosomes, histones, peptide lysine modifications, RNA splicing, ciliated epithelial tube morphogenesis, and acylation complex reactions involving proteins were upregulated; in contrast, signaling and histone acetylation mediated by small GTPases were down-regulated (Figure 1D). The molecular functions of KLF5 mRNA were activated by transcriptional regulators, chromatin binding, proximal promoter DNA binding, polymerase II proximal promoter sequence-specific DNA binding and enzyme activator activity, which were significantly upregulated, whereas small GTPase binding and histone acetylation were downregulated (Figure 1E). KEGG analysis of KLF5 revealed tumor-related pathways, such as adhesion, centrosome, nuclear dot, cell-substrate, and ciliary body pathways (Figure 1F). These data indicate that upregulated KLF5 could play important roles in promoting LUSC progression.

Figure 1 Bioinformatics analysis of lung Krüppel-like factor-5 mRNA and its function. A: KLF5 mRNA in lung adenocarcinoma (LUAD) (n = 585) or lung squamous-cell carcinoma (LUSC) (n = 550) tissues from the Cancer Genome Atlas (TCGA). database; B: KLF5 mRNA in LUSC or normal tissues from the Gene Expression Profiling Interactive Analysis (GEPIA) (Left) or TCGA (Right) databases; C: KLF5 mRNA in LUAD or normal tissues from the GEPIA (Left) or TCGA (Right) databases; D: Biological process of KLF5; E: Molecular function of KLF5; F: KEGG analysis of KLF5. KLF5: Krüppel-like factor 5 LUAD: Lung adenocarcinoma; LUSC: Lung squamous-cell carcinoma.

Upregulated KLF5 in tissues or sera of patients with lung cancer

A comparative analysis of KLF5 expression in the tissues and blood of patients with lung cancer is shown in Figure 2. KLF5 expression by MIF was located in the cytoplasm or nucleus and was increased in necrotic or infiltrating regions in LUSC (Figure 2A) and LUAD (Figure 2D) samples. KLF5 intensities in LUSC (Figure 2B) or LUAD (Figure 2E) were stronger than those in their adjacent tissues (Figure 2C or F). KLF5 Levels were confirmed at specific concentrations (KLF5 ng/mg wet lung) and were significantly increased in LUSC (P < 0.001, Figure 2C) but not in LUAD (P = 0.168, Figure 2F). The positive rate and intensity of KLF5 expression in lung tissues are summarized in Table 1. The incidence (85.0%, 51 of 60) of KLF5 or KLF5 intensity in LUSC was significantly different (χ² = 61.702, P < 0.001 or χ² = 74.278, P < 0.001) than that in Adj-Ca tissues. If the serum KLF5 concentration exceeded the upper limit of 500 ng/mL, the positive rates were 86.7% (52 of 60) in LUSC, 6.7% (4 of 60) in CP and 0% (0 of 60) in NC. The serum KLF5 Levels (F = 178.6, P < 0.001) significantly differed between the LUSC, CP and NC groups (Figure 2G), with abnormally upregulated KLF5 Levels in the sera of LUSC patients. The area under the receiver operating characteristic curve (ROC) of serum KLF5 was 0.85 (Figure 2H). These results revealed that the up-regulated KLF5 could be a useful diagnostic or differential diagnostic marker for LUSC.

Figure 2 Krüppel-like factor-5 expression in the lung tissues or sera of patients. A-C: Krüppel-like factor-5 (KLF5) in lung squamous-cell carcinoma (LUSC): B: KLF5 in Adj-Ca tissues of LUSC; C: KLF5 ng/mg wet lung tissues (n = 60), ^bP < 0.01; D-F: KLF5 in lung adenocarcinoma (LUAD): E: KLF5 in Adj-Ca tissues of LUAD; F: KLF5 ng/mg wet lung tissues (n = 60); G: The positive rates of serum KLF5 (> 800 ng/mL) in LUSC, chronic pneumonia and normal control, ^fP < 0.01; H: Receiver operating characteristic curve of circulating KLF5 Levels for LUSC diagnosis. Adj-Ca: Adjacent cancerous tissues; GAPDH: Glyceraldehyde-3-phosphate dehydrogenase; KLF5: Krüppel-like factor 5; LUAD: Lung adenocarcinoma; LUSC: Lung squamous-cell carcinoma; MIF: Multiplex immunofluorescence; LC: Lung cancer; CP: Chronic pneumonia; NC: Normal control; ROC: Receiver operating characteristic.

Table 1 Summary of the intensity of KLF5 expression in lung tissues.

Group	n	KLF5		χ2	P value	KLF5 intensity				Z value	P value
		Pos	Neg			-	+	++	+++
LUAD	60	39	21	2.749	0.097	21	29	9	1	5.059	0.168
Adj-Ca	60	29	31			31	19	10	0
LUSC	60	51	9	61.702	< 0.001	9	3	27	21	74.278	< 0.001
Adj-Ca	60	7	53			53	5	2	0

Adj-Ca: Adjacent cancerous tissues; KLF5: Krüppel-like factor 5; LUAD: Lung adenocarcinoma; LUSC: Lung squamous-cell carcinoma; Neg: Negative KLF5; Pos: Positive KLF5.

Clinicopathological characteristics of KLF5 in LUSC

The clinicopathological characteristics associated with positive KLF5 expression in LUSC patients are shown in Table 2. The clinical data of LUSC patients (n = 60) revealed that positive KLF5 expression in LUSC tissues were significantly correlated with poor differentiation (χ² = 8.989, P < 0.001) or lymph node metastasis (χ² = 6.424, P < 0.001). However, sex, age, tumor size, TNM stage or 5-year survival did not significantly differ between patients with KLF5 expression and LUSC patients. In addition, univariate/multivariate Cox regression (Table 3) revealed that positive KLF5 Levels were associated with poor differentiation (P < 0.001) and lymph node metastasis (P < 0.001) in LUSC patients. KLF5 was an oncogenic factor that could promote LUSC progression.

Table 2 Clinicopathological characteristics of positive Krüppel-like factor 5 expression in lung squamous-cell carcinoma.

Groups	No. of patient (%)	KLF5 positive (n)	%	χ2	P value
Lung cancer	60	51	85.0	-	-
Gender
Male	51	43	84.3	0.023	0.879
Female	9	8	88.9
Age (years)
≤ 60	28	23	82.1	0.047	0.828
> 60	32	28	87.5
Tumor diameter (cm)
≤ 2	15	12	80.0	0.044	0.835
> 2	45	39	86.7
Differentiation
Well	18	11	61.1	8.989	0.003a
Moderate/poor	42	40	95.2
Lymph node metastasis
Without	16	10	62.5	6.424	0.011a
With	44	41	93.2
TNM staging
I-II	49	41	83.7	0.020	0.889
III-IV	11	10	90.9
Five-year survival
Yes	19	17	89.5	0.074	0.786
No	41	34	82.9

^aP value < 0.05.

KLF5: Krüppel-like factor 5; TNM: Tumor node metastasis staging.

Table 3 Univariate/multivariate analysis of strong Krüppel-like factor 5 expression in lung squamous-cell carcinoma.

	Univariate			Multivariate
	HR	P > |z|	95%CI	HR	P > |z|	95%CI
KLF5 expression
Strong vs weak	3.050	0.002	1.741-5.326	3.122	0.003	1.725-5.458
Gender
Male vs female	1.704	0.052	0.994-2.921	1.121	0.706	0.622-2.015
Age (years)
≤ 60 vs > 60	1.115	0.628	0.717-1.733	1.365	0.826	0.845-2.209
Diameter (cm)
≤ 2 vs > 2	1.497	0.081	0.952-2.335	0.806	0.497	0.435-1.506
Differentiation
Well vs Mid/poor	3.216	< 0.001	0.886-3.021	2.012	< 0.001	0.065-2.688
LNM
No vs Mip/Mim	2.095	< 0.001	1.258-2987	2.095	< 0.001	1.197-3.667
Staging with TNM
I-II vs III/IV	0.925	0.642	0.432-1.586	1.515	0.082	0.948-2.422

HR: Hazard ratio; KLF5: Krüppel-like factor 5; LNM: Lymph node metastasis; Mid: Middle differentiation; Mim: Mediastinal metastasis; Mip: Ipsilateral peribronchial metastasis; No: No regional lymph node metastasis; TNM: Tumor node metastasis staging.

Effects of KLF5 on the biological behaviors of lung cancer cells

KLF5 expression in different lung cancer cell lines and the effects of KLF5 on biological behaviors are shown in Figure 3. Among the lung Beas-2B cells or lung cancer cell lines SPC-A-1, NCI-1650, A549 and NCI-1975, the levels of KLF5 expression were greater (P < 0.001) in SPC-A-1, NCI-1650, A549 and NCI-1975 cells than in Beas-2B cells (Figure 3A), as determined by Western blotting (Figure 3B). Stronger KLF5 expression was detected in A549 and NCI-H1975 cells based on the relative ratio of KLF5 to GAPDH (Figure 3C), and this expression was highly consistent with the expression of KLF5 mRNA (Figure 3D) according to quantitative real-time PCR. The effects of interfering with KLF5 transcription on biological behaviors were observed after specific KLF5i plasmids were transfected into A549 cells in vitro. Cell proliferation was significantly inhibited (Figure 3E), and KLF5 was down-regulated at the mRNA level (Figure 3F). The degree of cell migration (Figure 3G) significantly decreased (t = 9.209, P < 0.001) between the KLF5i (158.1 ± 29.6) and N-KLF5 (235.8 ± 21.8) groups. Additionally, the cell cycle was inhibited at the G1 phase and slight arrested at the G2 phase (Figure 3H), indicating that KLF5 transcription could be associated with the biological behaviors of lung cancer cells.

Figure 3 Intervening KLF5 affected the biological behaviors of lung cancer cells. A: Lung Beas-2B cells and lung cancer SPC-A-1, NCI-1650, A549 and NCI-1975 cells; B: Krüppel-like factor 5 (KLF5) at the protein level in different cells, as determined by Western blotting; C: Ratios of KLF5 to glyceraldehyde-3-phosphate dehydrogenase; D: KLF5 at the mRNA level in different cells, as determined by quantitative real-time PCR (qRT-PCR); E: Proliferating curve of A549 cells divided into blank, N-KLF5 or KLF5 intervention (KLF5i) groups and transfected with KLF5i plasmids; F: Alterations in KLF5 mRNA by qRT-PCR; G: Migrating analysis of A549 cells among different groups; H: Cell cycle alterations. GAPDH: Glyceraldehyde-3- phosphate dehydrogenase. ^aP < 0.05 and ^cP < 0.001, compared with the blank or N-KLF5 group. KLF5: Krüppel-like factor 5.

Effects of KLF5 intervention on xenograft tumor growth

The effects of KLF5 transcription interference on the growth of A549 nude mouse xenograft tumors are shown in Figure 4. After A549 cells transfected with KLF5i plasmids were inoculated subcutaneously into nude mice, the growth of xenograft tumors was significantly inhibited (Figure 4A). The average tumor weights (0.51 ± 0.18 g) in the KLF5i group were significantly lower (P < 0.001) than those in the N-KLF5 (1.51 ± 0.26 g) or Blank (1.49 ± 0.28 g) group (Figure 4B). Tumor growth was within the margin of error. The mean time of tumor formation in the KLF5i group (22.5 ± 1.2 days) was significantly longer (P < 0.001) than that in the N-KLF5 or Blank group (12.6 ± 0.8 days, Figure 4C). The intensities of KLF5 expression with brown or uniform granular staining in the xenograft tumors determined by IHC were greater in the Blank or N-KLF5 group (++ to +++, Figure 4D) and significantly lower (Z = 2.986, P < 0.001) in the KLF5i group (- to +, Figure 4E). Additionally, the MIF-induced increase in KLF5 expression in xenograft tumors was confirmed by stronger staining in the N-KLF5 or Blank group (Figure 4F) and weaker staining in the KLF5i group (Figure 4G). These results were verified by Western blotting, which showed a marked difference (P < 0.001) between the KLF5i group and the N-KLF5 or N-KLF5 group (Figure 4H). These data indicate that interfering with KLF5 transcription could be a promising target for inhibiting lung cancer growth.

Figure 4 Interfering with KLF5 transcription affected xenograft tumor growth. A: Nude mouse xenograft tumor growth after subcutaneous injection of A549 cells transfected with or without Krüppel-like factor 5 intervention (KLF5i) plasmids; B: Xenograft tumor weight (g); C: Xenograft tumor volume; D: Krüppel-like factor 5 (KLF5) straining by IHC in the N-KLF5 group (n = 6); E: KLF5 straining by IHC in the KLF5i group (n = 6); F: KLF5 straining by MIF in the N-KLF5 group; G: KLF5 straining by MIF in the KLF5i group; H: KLF5 in xenograft tumors by Western blotting (top) and their ratios from KLF5i to GAPDH (bottom). ^cP < 0.001, compared with the blank or N-KLF5 group. GAPDH: Glyceraldehyde-3-phosphate dehydrogenase; IHC: Immunohisto- chemistry; KLF5: Krüppel-like factor 5; MIF: Multiplex immunofluorescence.

KLF5-related pathways and mechanisms involved in promoting LUSC

The results of the enrichment analysis of KLF5 with related signaling pathways and possible mechanisms for promoting LUSC progression are shown in Figure 5. Biocarta analysis revealed that the KLF5 mRNAs were positively related to the Wnt, FAS, Mark and TGF-β signaling pathways and negatively correlated with AMI and the intrinsic pathway (Figure 5A-F). GO analysis revealed that the expression of the KLF5 mRNAs was positively correlated with cytokines, the apoptosis-executive phase, protein acylation, RNA splicing regulation, tight junction assembly and cytosolic mRNA splicing (Figure 5G-L). Based on the above analysis, a new mechanism of abnormal KLF5 transcription promoting LUSC progression through the activation of representative Wnt pathway was validated by human LUSC tissues and mouse xenograft tumors (Figure 5M). These data indicate that KLF5 cooperates with related signaling pathways to promote LUSC progression and that the oncogenic gene KLF5 could be a promising molecular target for LUSC therapy.

Figure 5 Krüppel-like factor 5-related pathways and possible mechanisms in lung cancer. A-F: Biocarta analysis: A: WNT_pathway; B: MAPK_pathway; C: FAS_pathway; D: TGFB_pathway; E: AMI_pathway; F: Intrinsic_ pathway; G-L: GO analysis: G: Cytokines; H: Execution_phase_ of_apoptosis; I: Protein_acetylation; J: Regulation of RNA splicing; K: Bicellular tight junction assembly; L: MRNA splicing by splicing exosomes; M: A possible new mechanism is that abnormal transcription of Krüppel-like factor 5 promotes lung squamous-cell carcinoma progression by activating related signaling pathways (such as Wnt and so on). KLF5: Krüppel-like factor 5; Wnt3a: A member of the Wnt signaling pathway.

DISCUSSION

The KLF family is a class of ZFP transcription factors that have oncogenic or anti-tumor effects and play extensive or critical roles in the proliferation and differentiation of tumors by targeting downstream genes[12,15]. Upregulated KLF expression has been observed in various cancers, including those of the pancreas, liver, and stomach[26-28], and the biological function of KLF may differ between these diseases. KLFs in cancers are divided into 4 categories: Promotion, inhibition, dual-function and unknown. Recently, increasing evidence has demonstrated that KLF5 can promote tumor progression in human cancers[14]. KLF5 gene is located at 13q21, spans 18.5 kb and contains 4 exons. KLF5 cDNA is 3350 bp and spans 324 bp at the 5’ end and 1652 bp at the 3’ end, and the DNA binding domain is located at the carboxyl end and contains three C2H2 zinc finger structures composed of highly conserved amino acids[29,30]. However, the role of KLF5 and its clinical value in LUSC remain unclear. This study investigated KLF5 transcription in LUSC via bioinformatic databases and verified this expression in clinical specimens. Furthermore, this study explored the effects of interfering with KLF5 gene transcription on the biological behavior of lung cancer cells or nude mouse xenograft tumor growth.

Lung carcinogenesis is a complex process that involves multiple pathogenic factors or steps[3,14], such as the activation of proto-oncogenes, the silencing of tumor-suppressor genes, and the inactivation of DNA repair genes and cell adhesion factors, which are key events in tumor biology[3,31]. LUAD and LUSC remain the most common types of lung cancer with different drivers of gene transcriptional status or epigenetic modifications, such as histone modifications and miRNAs[2,4]. Interestingly, upregulated KLF5 mRNAs were identified in LUSC and LUAD tissues via the GEPIA or TCGA databases. KLF5 functions are characterized by the activities of transcriptional regulators, chromatin, proximal promoter sequence-specific DNA and polymerase II proximal promoter sequence-specific DNA binding[32]. However, small GTPase binding and histone acetylation are down-regulated. KLF5 is related to signaling pathways such as adhesion junctions, centrosomes, nuclear dots, cell–substrate junctions, ciliary bodies, focal adhesion, Wnt, FAS, MARK and TGF-β, but not AMI or the intrinsic pathway[33,34]. However, significant differences between cancer and noncancerous tissues were identified, with high KLF5 mRNA expression in LUSC. Taking the above lung cancer databases into account, screening results need to be further validated with clinical samples[35].

KLF5 has attracted attention because of its important regulatory activities linked to diverse functions[14]. Increasing evidence indicates that KLF5 is a vital component in promoting tumor progression[36]. Stronger KLF5 mRNA transcription in LUSC tissues was validated at the protein level. First, strong KLF5 staining was confirmed by MIF or Western blotting, and KLF5 was quantified at a tissue-specific concentration in LUSC. The positive rate or intensity of KLF5 in LUSC was significantly greater than that in Adj-Ca tissues, but this difference was not observed in LUAD tissues. Second, the clinical features of stronger KLF5 expression in LUSC were significantly correlated with poor differentiation or lymph node metastasis. No significant differences were found between patients with high KLF5 expression and LUSC patients in terms of sex, age, tumor size, TNM stage or 5-year survival. Moreover, a Cox regression analysis of KLF5 revealed that poor differentiation and lymph node metastasis were independent risk factors for LUSC patients. Moreover, serum KLF5 Levels in LUSC were verified that should be associated with tissue KLF5 upregulation. Additionally, the area under the ROC curve of KLF5 in LUSC patients was significantly greater than that in cases with benign lung diseases or healthy individuals. These data indicate that high KLF5 at mRNA or protein level could be a useful biomarker for LUSC diagnosis or differential diagnosis[37,38].

Although studies on upstream KLF5 have been reported in other cancers[39], reports on how it regulates lung cancer progression are scarce[14,40]. An analysis of the correlations between KLF5 and pathological characteristics revealed that KLF5 expression generally increased with tumor growth and nodal metastasis in LUSC. All lung cancer cells expressed abnormal levels of KLF5, and this expression positively correlated with biomarkers of LUSC infiltration. Stronger KLF5 expression promoted LUSC proliferation, migration and tumor growth in vitro. Recently, KLF5 was reported to promote migration in laryngeal squamous cell carcinoma. Interestingly, interfering with KLF5 transcription affected the biological behaviors of A549 cells and nude mouse xenograft tumors in vivo. Moreover, KLF5 promoted tumorigenesis via the activation of related pathways such as Wnt in liver[41,42] or lung cancer[43], which was confirmed in the tissues of human LUSC or xenograft tumors. The activation of the KLF5 and the Wnt signaling pathway could be a novel mechanism underlying LUSC progression and could provide a promising molecular target for LUSC therapy[44]. Therefore, further investigations of the possibility of targeting KLFs in LUSC therapy are urgently needed

CONCLUSION

Our data provide some novel evidences to show that KLF5 is upregulated in lung cancer, with stronger expression in LUSC than in LUAD. Higher KLF5 expression regulated the proliferation, migration and cell cycle of LUSC, which is likely a favorable diagnostic biomarker of LUSC or a new molecular target for LUSC. Furthermore, KLF5 might exert its oncogenic regulatory effects on the growth of LUSC by increasing metastasis or LUSC progression by activating related signaling pathways. However, the exact mechanisms involved need to be confirmed with more basic and clinical studies. Similarly, strategies to silence KLF5 transcription plus Wnt signaling or multi-targeting strategies need to be investigated for effective LUSC therapy.

ACKNOWLEDGEMENTS

We are thankful to Professor Li Wang for his critical and careful editing and proofreading of the manuscript. Authors thank the staff of the Research Center of Medical Research, Affiliated Hospital of Nantong University, China.