Gene Expression Omnibus (GEO) is a publicly available database that provides gene expression data for various organisms, including humans. In this study, the GEO dataset GSE62663 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE62663) was analyzed using GEO2R to examine the expression levels of KLF8 in LUAD cells under different culture conditions [22]. Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia.cancer-pku.cn/) is a web-based tool that allows users to analyze gene expression data from the Cancer Genome Atlas (TCGA, http://cancergenome.nih.gov/) and the Genotype-Tissue Expression (GTEx, https://www.gtexportal.org/home/) project. In this study, GEPIA was used to assess the correlation between KLF8 and the histone acetyltransferase KAT9 and the deacetylase SIRT2 in LUAD. Human Transcription Factor Database (HumanTFDB, http://bioinfo.life.hust.edu.cn/HumanTFDB#!/) is a comprehensive database that provides information regarding transcription factors and their binding sites. In this study, the HumanTFDB website was used to predict the binding sites of KLF8 and SIRT2 within the promoter regions of their target genes.

Lung adenocarcinoma tissues (n = 30) and adjacent normal tissues (n = 30) were collected from patients with LUAD in our hospital. Samples were immediately stored at –80 °C for further analysis. The Ethics Committee of our hospital has approved this study, and all patients who participated in this study have signed a written informed consent.

BEAS-2B cells (95102433, ECACC, UK) were cultured in Bronchial Epithelial Cell Growth Medium (BEGM, CC-3171, Lonza, Switzerland). A549, H1299, HCC827 and H460 cells (CCL-185; CRL-5803; CRL-2868; HTB-177, ATCC, Manassas, VA, USA) were incubated in RPMI-1640 (11875093, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) with 10% fetal bovine serum (FBS, 10270-106, Gibco, Grand Island, NY, USA). Prior to experimentation, the cells were allowed to attach overnight in a humidified incubator at 37 °C with 5% CO${}_2$. The next day, nutrient-deprived medium was added for an additional 48 h incubation, as indicated. The nutrient-deprived medium was prepared by using BEGM without the addition of FBS.

A Mycoplasma PCR Detection Kit (AAJ66117AMJ, Thermo Fisher Scientific) was used for screening for potential mycoplasma infection. Cells have been authenticated by short tandem repeat (STR) profiling, which are listed by corresponding manufacturers on their product websites.

The short hairpin RNAs (shRNAs) targeting KLF8 (sh-KLF8, 5${}^{\prime }$-CAGCACTGTTTAATGACAT-3${}^{\prime }$) and a corresponding negative (non-targeting) control (sh-NC, 5${}^{\prime }$-TTCTCCGAACGTGTCACGT-3${}^{\prime }$) were designed and synthesized into the shRNA expression vector (pGPU6/mCherry/Puro) by GenePharma Co. Ltd. (Shanghai, China). To overexpress FLF8 and G6PD, pcDNA3.1/KLF8, and pcDNA3.1/G6PD vectors were constructed and provided by the Genomeditech Shanghai Co., Ltd. (Shanghai), and empty pcDNA3.1 vector was used as the negative control. H1299 and A549 cells were seeded into 6-well plates at 3 $\times$ 10${}^5$ cells/well and transfected with the above vectors or negative control vectors using Lipofectamine 3000 (Invitrogen, Waltham, MA, USA). The stably transfected cells were selected using puromycin [23].

Protein extracts were prepared using Radio Immunoprecipitation Assay (RIPA) lysis buffer (Thermo Fisher Scientific) and protein concentrations measured using a Bicinchoninic acid (BCA) protein assay kit (23225, Thermo Fisher Scientific). Proteins were subsequently separated by SDS-PAGE and transfered to (polyvinylidene difluoride) PVDF membranes (IPVH00010, Millipore, Burlington, MA, USA). Following this, membranes were cultured with anti-KLF8 (1:1000, ab168527, Abcam), G6PD (1:1000, ab210702, Abcam, Cambridge, UK), anti-SIRT2 (1:1000, 12640, Cell Signaling, Danvers, MA, USA) and anti-$\beta$-actin (1:1000, AF0003, Beyotime, Shanghai, China) at 4 °C overnight. After washing the membranes, secondary antibodies (A0208, 1:1000, Beyotime) were added for a 1 h incubation at room temperature. Enhanced chemiluminescence reagent (P0018, Beyotime) and ImageJ2 software (LOCI, University of Wisconsin, Madison, WI, USA) were used for visualization and quantification of proteins.

After deparaffinization and rehydration, tissue sections from LUAD patients were incubated with anti-KLF8 (1/200, ab137676, Abcam, UK) primary antibody overnight at 4 °C. This step was followed by incubation with appropriate HRP-conjugated secondary antibody (1/5000, ab97051, Abcam). The slides were visualized using DAB (3,3${}^{\prime }$-Diaminobenzidine) substrate (ab94665, Abcam) and counterstained with hematoxylin. Images were captured using a light microscope (Olympus, Tokyo, Japan).

PrimeScript RT reagent kit (RR037A, TaKaRa, Beijing, China) was used for cDNA synthesis. Quantitative real-time PCR (qRT-PCR) was performed using a LightCycler 480 Real-Time PCR System (Roche Diagnostics, Rotkreuz, Switzerland) and SYBR Premix Ex Taq II (RR820A, TaKaRa). The primers used in qRT-PCR were as follows: KLF8(Forward): 5${}^{\prime }$- TATTCAAGCAGGTCACTGC-3${}^{\prime }$, KLF8(Reverse): 5${}^{\prime }$-CAGGAGTGTTGGAGAAGTC-3${}^{\prime }$; G6PD(Forward): 5${}^{\prime }$-GTGAGGACCAGATCTACCG-3${}^{\prime }$, G6PD(Reverse): 5${}^{\prime }$-TGTTGGCAAATCTCAGCAC-3${}^{\prime }$; SIRT2(Forward): 5${}^{\prime }$-AGCTGGAATCTCCACATCC-3${}^{\prime }$, SIRT2(Reverse): 5${}^{\prime }$-CCGGATGTTTCTTGAAATAGCT-3${}^{\prime }$; GAPDH(Forward): 5${}^{\prime }$-AGCCACATCGCTCAGACAC-3${}^{\prime }$, GAPDH(Reverse): 5${}^{\prime }$-GCCCAATACGACCAAATCC-3${}^{\prime }$.

Nighty-six-well plates were inoculated at 5000 cells/well. Forty-eight hours later, 10 µL of Cell counting kit-8 (CCK-8) reagent (CK04-11, Dojindo, Kyushu Island, Japan) was added to each well. After a 2 hr incubation at 37 °C, absorbance at 450 nm was recorded using a BioTek Epoch microplate reader (Agilent, Beijing, China).

Six-well plates were inoculated at 500 cells/well, with medium refreshed every 3 days. Paraformaldehyde (4%) was used to fix the cell monolayer for 15 min and crystal violet solution (0.1%, C0775, Sigma-Aldrich, St. Louis, MO, USA) was used to stain the colonies for 30 min. Excess stain was rinsed off with water, and the plates were air-dried. The number of visible colonies, defined as clusters of at least 50 cells, was counted manually under a light microscope (Olympus, Japan).

After washing with pre-chilled PBS, H1299 and A549 cells (2 $\times$ 10${}^5$ cells) were collected by centrifugation for 5 min at 1500 g. Then cells were then resuspended in binding buffer containing 5 µL Annexin-V-FITC and 5 µL propidium iodide (PI) and incubated for 20 min at room temperature in dark. Cell staining was analyzed using a flow cytometer (BD Biosciences, San Jose, CA, USA). With AnnexinV staining set on the horizontal axis and PI staining as the vertical axis, the apoptotic cells were quantified using CellQuest software (version 5.1, BD Biosciences).

To assess ROS in H1299 and A549 cells, a fluorescence assay was performed using 2${}^{\prime }$,7${}^{\prime }$-dichlorofluorescein diacetate (DCFH-DA) probe (Beyotime). Briefly, 96-well plates were inoculated with LUAD cells at 5 $\times$ 10${}^3$ cells/well. Then cells were washed and treated with 10 µM DCFH-DA (D6883, Sigma-Aldrich) at 37 °C for 30 min. The fluorescence intensity of 2${}^{\prime }$,7${}^{\prime }$-dichlorofluorescein, indicative of ROS levels, was analyzed using a fluorescence microplate reader (PerkinElmer, Waltham, MA, USA).

The levels of NADP${}^{+}$/NADPH in RIPA protein extracts were assayed based on the enzymatic cycling reaction using a commercial kit (ab65349, Abcam). Briefly, co-incubation of cell lysates, reaction buffer, and NADPH cycling enzyme mix was conducted for 30 min and results were obtained using a spectrophotometer at an absorbance of 450 nm.

Oxidative PPP flux was determined by measuring ${}^{14}$CO${}_2$ release as previously indicated [24]. LUAD cells were incubated in a 6-cm dish within a 10-cm dish containing two sealed pinholes on top. Cells were exposed to 2 mL of medium containing [6-${}^{14}$C]-glucose or [1-${}^{14}$C]-glucose for 3 h at 37 °C. To halt the PPP flux, 0.3 mL of 50% TCA was injected through one pinhole, with 0.3 mL of hyamine hydroxide injected through the second pinhole at the same time. Following this, both dishes were sealed and incubated overnight at room temperature. Hyamine hydroxide was dissolved in 60% methanol and subsequently analyzed via scintillation counting (Beckman Coulter, LS6500 Scintillation Counter, GA, USA).

H1299 and A549 cells were seeded in culture plates and allowed to adhere overnight. Cells were then treated with 1 mM 5-ethynyl uridine (EU) (E10345, Thermo Fisher Scientific) for 1 h. Next, cells were harvested, and total RNA extracted using a QIAGEN RNeasy Mini Kit (74104, QIAGEN, Hilden, Germany). The EU incorporated into RNA was then labeled using an EU detection kit (C10329, Click-iT RNA Imaging Kit, Thermo Fisher Scientific). Labeled RNA was visualized using fluorescence microscopy (Leica, Wetzlar, Germany), and the intensity of the fluorescence signal was quantified using ImageJ2. The fluorescence intensity directly corresponds to the level of nascently synthesized RNA in the cells.

H1299 and A549 cells were lysed and centrifuged, and the supernatant containing the protein extracts was collected. To measure G6PD activity, a G6PD assay kit (ab102529, Abcam, UK) was used. The reaction mixture, consisting of cell lysate, reaction buffer, and NADP${}^{+}$ cycling enzyme mix, was incubated for 30 min at 37 °C. The enzymatic reaction catalyzed by G6PD results in the reduction NADP${}^{+}$/NADPH, which was visualized by a change in absorbance. The absorbance was measured at 450 nm using a BioTek Epoch microplate reader. The rate of NADPH production, indicated by the change in absorbance over time, was used as a measure of G6PD enzyme activity.

Briefly, H1299 and A549 cells were lysed and precleared with protein A/G beads (sc-2003, Santa Cruz Biotechnology, Shanghai, China). Anti-G6PD antibody (ab993, Abcam) was added for incubation overnight at 4 °C. The immunocomplexes were collected with protein A/G beads and washed with lysis buffer. The beads were then eluted with 2$\times$ SDS sample buffer (1610737, Bio-Rad, Hercules, CA, USA) and the eluted proteins were analyzed via western blotting using anti-acetylated-Lysine antibody (9441, Cell Signaling) to detect the acetylation of G6PD. The signals were detected using enhanced chemiluminescence (34095, Thermo Fisher Scientific).

The binding sites of KLF8 on SIRT2 promoters were predicted by the HumanTFDB website. The SIRT2 fragments containing wild-type (WT) or mutant (Mut) binding sequences were subcloned into the pGL3-basic luciferase reporter vectors to construct promoter-WT or promoter-Mut plasmids. Subsequently, H1299 and A549 cells were co-transfected with promoter-WT or promoter-Mut constructs along with sh-NC or sh-KLF8 using Lipofectamine 2000 (Thermo Fisher Scientific). After 48 h, the cells were lysed using Passive Lysis Buffer (E1941, Promega, USA). Luciferase activity was measured using a Dual-Luciferase Reporter Assay System (E1910, Promega, Madison, WI, USA). Relative Firefly luciferase activity was normalized to Renilla luciferase as the internal standard. The relative luciferase activity was calculated by comparing the promoter-WT or promoter-Mut transfected with either the negative (sh-NC) or positive (sh-KLF8) control groups.

H1299 and A549 cells were treated with 1% formaldehyde (Sigma-Aldrich) to cross-link DNA and proteins. Following this, the chromatin was sonicated to generate chromatin fragments using a Branson digital sonifier SFX 550 (Emerson, Danbury, CT, USA). Immunoprecipitation was carried out using anti-KLF8 (PA5-109885, 1:1000, Thermo Fisher Scientific), with a non-specific IgG (ab6715, 1:10,000, Abcam) used as the control. The immunoprecipitated DNA was purified and analyzed by qRT-PCR. Amplified PCR products were loaded on a 2% agarose gel with ethidium bromide and ImageJ2 software was used for quantitation. Relative enrichment of SIRT2 promoter fragment binding to anti-KLF8 or anti-IgG was calculated by comparing the ChIP signal intensity to the input DNA control [25]. The primer sequences used in the ChIP assays were as follows: SIRT2: Forward: 5${}^{\prime }$-CACCACATGGCTGTCTGTTT-3${}^{\prime }$; Reverse: 5${}^{\prime }$-TGGTGGAGGTGGCTAAATGT-3${}^{\prime }$.

Data were processed using GraphPad Prism software (version 8.0.1, GraphPad Software Inc., San Diego, CA, USA) and presented as the mean $\pm$ standard deviation (SD). All experiments were conducted in triplicate. Student’s t-test was used to analyze the statistical difference between 2 groups, while one-way ANOVA followed by Tukey’s test was used for comparisons among three or more groups. Correlation was analyzed using Pearson’s correlation coefficient. Each experiment was performed in triplicate. A p-value of less than 0.05 was considered statistically significant.

The expression of KLF8 under serum starvation conditions was initially examined to investigate its role in nutrient deprivation. qRT-PCR, western blotting (WB) and immunohistochemistry (IHC) revealed upregulation of both relative KLF8 transcript and protein expression in LUAD tissues, when compared to normal controls (Fig. 1A–C). Additionally, qRT-PCR demonstrated higher KLF8 mRNA abundance in numerous LUAD cell lines, specifically, A549, H460, H1299, and HCC827, when compared to BEAS-2B normal cells (Fig. 1D). The expression of KLF8 in H3122 cells in serum-containing and serum starvation conditions was analyzed by GEO2R based on GSE62663 dataset, and the results revealed that KLF8 expression was significantly upregulated in LUAD cells under serum starvation conditions compared to control conditions (Fig. 1E). qRT-PCR and WB experiments also demonstrated higher KLF8 expression in LUAD cells under serum deprivation than under normal culture medium serum conditions (Fig. 1F–G).

Fig. 1.

Krüppel-like factor 8 (KLF8) is upregulated in lung adenocarcinoma (LUAD) tissues and serum starved LUAD cell lines. (A) Quantitative real-time polymerase chain reaction (qRT-PCR) was used to measure the expression of KLF8 in tumor tissues (n = 30) and adjacent normal tissues (n = 30). (B,C) KLF8 expression in normal/LUAD tissues (n = 5), detected via western blotting (WB) and immunohistochemistry (IHC). (D) KLF8 expression in normal/LUAD cells (A549, H460, H1299, HCC827) and human bronchial epithelial cells (BEAS-2B), detected via qRT-PCR. (E) KLF8 expression under starvation or control conditions, analyzed in the GSE62663 dataset using GEO2R. (F,G) Relative expression of KLF8, detected via qRT-PCR or WB, in H1299 and A549 cells grown in serum (+) or no serum (-) groups. *p 0.05, **p 0.01, ***p 0.001.

A previous study has reported that KLF8 promotes lung cancer cell growth and cell cycle progression in normal culture conditions [18]. Thus, we focused on the role of KLF8 in LUAD cell survival under nutrient-deprived conditions via overexpression and RNA interference experiments. Relative overexpression and interference efficiency are shown in Fig. 2A,B. CCK-8 assays showed that the LUAD cell viability was suppressed under serum starvation conditions compared with the serum-fed condition. Additionally, H1299 and A549 cell viability was relatively higher in the recombinant KLF8 serum-deprivation group relative to the Vector only serum-deprivation group, and further inhibited in the KLF8 knockdown (sh-KLF8 and serum-deprivation group relative to the shRNA control, serum-deprivation group (Fig. 2C,D). Colony formation assays revealed that overexpression of KLF8 increased the number of H1299 and A549 cell colonies in the KLF8 serum-deprivation group compared to the Vector serum-deprivation group, while KLF8 knockdown decreased colony formation in the KLF8-knockdown, serum (-) group compared to the control knockdown (sh-NC) serum-deprivation group (Fig. 2E,F). Flow cytometry indicated that KLF8 overexpression suppressed apoptosis in the KLF8 serum-deprivation group compared to the Vector control, serum (-) group, while KLF8 knockdown elevated apoptosis in the sh-KLF8 serum (-) group compared to the control sh-NC serum (-) group (Fig. 2G,H).

Fig. 2.

KLF8 promoted the survival of H1299 and A549 cells under serum starvation conditions. (A,B) KLF8 overexpression and shRNA knockdown efficiency in H1299 and A549 under serum starvation conditions as assessed by qRT-PCR. ***p 0.001 when compared to the control sh-NC vector group. (C,D) Cell viability of H1299 and A549 cells as assayed via cell counting kit-8 (CCK-8). (E,F) Cell growth of H1299 and A549 cells, assessed via colony formation assay. (G,H) Apoptosis of H1299 and A549 cells with KLF8 overexpression and knockdown and cultured in the presence or absence of serum was assessed via flow cytometry. *p 0.05, **p 0.01, ***p 0.001 vs the serum control (+) group; ${}^{\mathrm{\#}}$p 0.05, ${}^{\mathrm{\#}\mathrm{\#}}$p 0.01, ${}^{\mathrm{\#}\mathrm{\#}\mathrm{\#}}$p 0.001 vs the control sh-NC serum (-) or Vector serum (-) group.

It has been reported that under high ROS conditions, tumors can maintain cell proliferation by regulating metabolic reprogramming through PPP [26], and some studies have found that modulating PPP under cell starvation conditions contributes to cell survival [10]. We found that the ROS level was significantly increased in H1299 and A549 cells during nutrient deprivation (serum (-)) conditions compared with the control sh-NC serum (+) or Vector serum (+) groups. This was further enhanced by KLF8 knockdown and decreased by recombinant KLF8 overexpression. Moreover, overexpression or knockdown of KLF8 in serum (+) condition showed no significant impact on the ROS levels in H1299 and A549 cells (Fig. 3A,B).

Fig. 3.

KLF8 promoted the metabolism in H1299 and A549 cells via pentose phosphate pathway (PPP). (A,B) Relative reactive oxygen species (ROS) level in H1299 and A549 cells from sh-NC/KLF8 or Vector/KLF8 group and cultured under serum (+) or serum (-) conditions. (C,D) NADPH/NADP${}^{+}$ ratio in H1299 and A549 cells from sh-NC/KLF8 or Vector/KLF8 group and cultured in serum (+) or serum (-) medium. (E,F) Relative oxidate PPP flux (%) in H1299 and A549 cells transfected with sh-NC/KLF8 or Vector/KLF8 and cultured in serum (+) or serum (-) medium. (G,H) The effects of KLF8 knockdown or overexpression on RNA biosynthesis rates in H1299 and A549 cells cultured in serum (+) or serum (-) medium. *p 0.05, **p 0.01, ***p 0.001 vs the sh-NC/Vector serum (+) group; ${}^{\mathrm{\#}}$p 0.05, ${}^{\mathrm{\#}\mathrm{\#}}$p 0.01, ${}^{\mathrm{\#}\mathrm{\#}\mathrm{\#}}$p 0.001 vs the sh-NC serum (-) or Vector serum (-) group.

We next explored whether KLF8 had an effect on PPP. KLF8 knockdown (Fig. 3C,E,G) resulted in significantly lower NADPH/NADP${}^{+}$ ratios, decreased oxidate PPP flux, and lower RNA synthesis rates when compared to the control sh-NC/serum (-) group and the sh-NC serum (+) group. Levels were further elevated by KLF8 silencing in serum starvation conditions compared with the sh-NC serum (-) group. KLF8 overexpression (Fig. 3D,F,H) showed that compared with the Vector serum (+) group, the Vector serum (-) group had significantly higher NADPH/NADP${}^{+}$ ratios, increased oxidate PPP flux, and higher RNA synthesis rates. Further, these levels were reduced by KLF8 overexpression in serum (-) conditions compared with Vector serum (-) group. Moreover, we also observed that KLF8 overexpression or knockdown in serum (+) conditions did not significantly change the NADPH/NADP${}^{+}$ ratios, oxidate PPP flux and RNA synthesis rates compared with the control serum (+) groups. These results suggest that KLF8 activated the PPP and promoted increased metabolism in H1299 and A549 cells under starvation conditions.

ShRNA-mediated knockdown of KLF8 significantly decreased G6PD activity in both H1299 and A549 cell lines compared to the sh-NC control group under both serum (+) and serum (-) conditions, and the overexpression of KLF8 had the opposite effect (Fig. 4A,B). However, qRT-PCR and western blotting confirmed that sh-KLF8 or KLF8 overexpression had no significant impact on G6PD expression at either the mRNA and protein levels in H1299 and A549 cells in both serum (+) and serum (-) conditions (Fig. 4C–F). In addition, LUAD patients with high G6PD expression exhibited poor overall survival (p = 4.4 $\times$ 10${}^{-7}$) (Fig. 4G). It has been reported that G6PD activity is affected by protein acetylation [27].

Fig. 4.

KLF8 inhibits glucose-6-phosphate dehydrogenase (G6PD) acetylation to elevate G6PD protein activity. (A,B) Relative G6PD activity in H1299 and A549 cells under serum (+) and serum (-) conditions was measured in sh-NC or sh-KLF8 or Vector or KLF8-overexpression groups. (C–F) Relative G6PD expression in H1299 and A549 cells under serum (+) and serum (-) conditions in sh-NC or sh-KLF8 or Vector or KLF8-overexpression groups, assessed via qRT-PCR and WB. (G) Survival plot of LUAD patients with high or low G6PD expression using the Kaplan-Meier Plotter database. (H,I) Acetylation of G6PD in H1299 and A549 cells under serum (+) and serum (-) conditions from sh-NC and sh-KLF8 groups, assessed by immunoprecipitation/WB. **p 0.01, ***p 0.001.

To better understand this effect, acetylation of the G6PD protein was investigated in H1299 and A549 cells transfected with sh-NC or sh-KLF8 using immunoprecipitation. Using anti-G6PD, pan-acetylation antibody detected increased lysine acetylation in the sh-KLF8 group under both serum (+) and serum (-) conditions. This finding demonstrated that KLF8 knockdown increased the acetylation of G6PD in both H1299 and A549 cells (Fig. 4H,I). Further, these findings suggested that KLF8 enhances G6PD activity through promoting G6PD post-translational modification rather than controlling transcriptional regulation during both serum (+) and serum (-) conditions.

To investigate whether G6PD mediated the effect of KLF8 on serum starvation-induced growth inhibition, H1299 and A549 cells were transfected with G6PD under serum (-) condition, overexpression efficiency is shown in Fig. 5A. As shown in Fig. 5B, overexpression of G6PD partially rescued the cell viability reduction caused by KLF8 knockdown in both H1299 and A549 cells under serum starvation. The colony formation assays further confirmed that G6PD overexpression partially reversed the inhibitory effect of sh-KLF8 on LUAD cell growth under serum (-) conditions (Fig. 5C). Moreover, flow cytometry analysis revealed that G6PD overexpression significantly decreased apoptosis rates induced by KLF8 knockdown in both cell lines under serum starvation conditions (Fig. 5D). Furthermore, G6PD overexpression also reversed the decrease of measured ROS levels (Fig. 5E), NADPH/NADP${}^{+}$ ratio (Fig. 5F), oxidate PPP flux (Fig. 5G), and RNA synthesis (Fig. 5H) were each induced by KLF8 knockdown in H1299 and A549 cells during serum starvation. In addition, we also verified that KLF8 reversed the serum starvation-induced growth inhibition via G6PD using the G6PD inhibitor 6-aminonicotinamide (6-AN). As shown in Fig. 6A,B, the viability and proliferation of H1299 and A549 cells was elevated by KLF8 overexpression under serum (-) condition compared with the vector serum (-) group, and the 6-AN administration restored the basal level of proliferation. Moreover, apoptosis in H1299 and A549 cells was reduced by overexpressing KLF8 under serum (-) conditions relative to the Vector serum (-) group. Furthermore, 6-AN treatment rescued the impact that KLF8 overexpression exerted on LUAD cell apoptosis (Fig. 6C). Additionally, 6-AN treatment reversed ROS levels, NADPH/NADP${}^{+}$ ratio, oxidate PPP flux, and RNA synthesis rate increased as a result of KLF8 overexpression in H1299 and A549 cells under serum (-) conditions (Fig. 6D–G). These results suggest that KLF8-regulated, serum starvation-induced, growth inhibition occurs through promoting G6PD-mediated PPP activity.

Fig. 5.

KLF8 regulates the serum starvation-induced inhibition on cell growth via G6PD. (A) Overexpression of G6PD in H1299 and A549 cells as measured via qRT-PCR under serum (-) conditions. (B) Cell viability of H1299 and A549 cells under serum (-) conditions from sh-NC/sh-KLF8/sh-KLF8+G6PD group as assessed by CCK-8. (C) H1299 and A549 cell growth under serum (-) conditions from sh-NC/sh-KLF8/sh-KLF8+G6PD groups as assessed by colony formation assay. (D) Apoptosis of H1299 and A549 cells under serum (-) conditions from sh-NC/sh-KLF8/sh-KLF8+G6PD group as assessed by flow cytometry. (E) Relative ROS level in H1299 and A549 cells under serum (-) condition from sh-NC/sh-KLF8/sh-KLF8+G6PD group cells as assessed using fluorescent probes. (F) NADPH/NADP${}^{+}$ ratio in H1299 and A549 cells under serum (-) condition from sh-NC/sh-KLF8/sh-KLF8+G6PD group cells as measured using a colorimetric assay. (G) Relative oxidate PPP flux (%) in H1299 and A549 cells under serum (-) condition from sh-NC/sh-KLF8/sh-KLF8+G6PD group cells. (H) Relative RNA synthesis rates under serum (-) conditions in H1299 and A549 cells from sh-NC/sh-KLF8/sh-KLF8+G6PD group cells. ***p 0.001 when compared to the sh-NC serum (-) group; ${}^{\mathrm{\#}\mathrm{\#}}$p 0.01, ${}^{\mathrm{\#}\mathrm{\#}\mathrm{\#}}$p 0.001 when compared to the sh-KLF8 serum (-) group.

Fig. 6.

The 6-AN treatment reversed the KLF8-induced enhancement on cell survival under serum starvation conditions. (A) Cell viability of H1299 and A549 cells under serum (-) conditions in the Vector/KLF8/and KLF8+6-AN group as measured by CCK-8 assay. (B) H1299 and A549 cell growth under serum (-) conditions in the Vector/KLF8/KLF8+6-AN groups as assessed by colony formation assay. (C) Apoptosis of H1299 and A549 cells under serum (-) conditions in Vector/KLF8/KLF8+6-AN group as measured by flow cytometry. (D) Relative ROS level in H1299 and A549 cells under serum (-) conditions in Vector/KLF8/KLF8+6-AN groups measured using fluorescent probes. (E) NADPH/NADP${}^{+}$ ratio in H1299 and A549 cells under serum (-) conditions in Vector/KLF8/KLF8+6-AN groups measured using a colorimetric assay. (F) Relative oxidate PPP flux (%) in H1299 and A549 cells under serum (-) conditions in Vector/KLF8/KLF8+6-AN groups. (G) Relative RNA synthesis rate under serum (-) conditions in H1299 and A549 cells in Vector/KLF8/KLF8+6-AN groups. *p 0.05, **p 0.01, ***p 0.001 when compared to the Vector serum (-) group; ${}^{\mathrm{\#}}$p 0.05, ${}^{\mathrm{\#}\mathrm{\#}}$p 0.01, ${}^{\mathrm{\#}\mathrm{\#}\mathrm{\#}}$p 0.001 when compared to the KLF8 serum (-) group.

Based on previous findings [20], KAT9 and SIRT2 are known to regulate G6PD acetylation. Gene Expression Profiling Interactive Analysis (GEPIA) analysis showed that KLF8 and SIRT2 were positively correlated in lung cancer tissues, and that there was no significant correlation between the expression of KLF8 and KAT9 (Fig. 7A). We next examined the effects of KLF8 knockdown or overexpression on the expression of KAT9 and SIRT2 using qRT-PCR and WB in A549 and H1299 cells transfected with sh-NC, sh-KLF8, empty Vector, or KLF8-expressing Vector. The results showed a positive correlation between SIRT2 and KLF8 expression (Fig. 7B–D). Moreover, the Kaplan-Meier Plotter website showed that LUAD patients with high SIRT2 expression had adverse survival outcomes (Fig. 7E).

Fig. 7.

KLF8 acted as a transcription factor of SIRT2. (A) Correlation between KLF8 and SIRT2, KLF8 and KAT9 was analyzed using Gene Expression Profiling Interactive Analysis (GEPIA). (B–D) Relative expression of KAT9 and SIRT2 in H1299 and A549 cells in the sh-NC/KLF8 or Vector/KLF8 group, detected by qRT-PCR and WB. (E) Survival plot of LUAD patients with high or low expression of SIRT2 on the Kaplan-Meier Plotter website. (F) Binding sites of KLF8 on SIRT2 promoters were predicted by the HumanTFDB website. (G) Relative luciferase activity in A549 and H1299 cells transfected with sh-NC/KLF8 and promoter-WT/Mut of SIRT2. (H) Relative enrichment (% of input) of SIRT2 promoter in the precipitates of anti-IgG/KLF8 in H1299 and A549 cells was detected using chromatin immunoprecipitation (ChIP) assays. (I) Graphical abstract of the study. ***p 0.001.

Next, the binding sites for KLF8 on the SIRT2 promoter was predicted using the HumanTFDB website (Fig. 7F). To confirm the results of this analysis, luciferase activity in H1299 and A549 cells showed a significant decrease after KLF8 knockdown in the group transfected with the promoter-WT probe. In contrast, findings showed no significant alteration in luciferase activity when cells were transfected with promoter-MUT. These findings indicate that KLF8 activates SIRT2 transcription (Fig. 7G). Finally, chromatin immunoprecipitation (ChIP) results indicated that the SIRT2 promoter showed higher enrichment after immunoprecipitation with anti-KLF8 compared with negative control anti-IgG. This result indicates that KLF8 directly interacts with the SIRT2 promoter (Fig. 7H) and suggest that KLF8 may promote H1299 and A549 cell growth and survival by activating SIRT2 transcription (Fig. 7I).