Abstract

   Patients with lung adenocarcinoma (LUAD) generally have poor prognosis.
   Abnormal cellular energy metabolism is a hallmark of LUAD.
   Glutathione-specific gamma-glutamylcyclotransferase 1 (CHAC1) is a
   member of the γ-glutamylcyclotransferase family and an unfolded protein
   response pathway regulatory gene. Its biological function and molecular
   regulatory mechanism, especially regarding energy metabolism underlying
   LUAD, remain unclear. By utilizing tissue microarray and data from The
   Cancer Genome Atlas and Gene Expression Omnibus, we found that CHAC1
   expression was markedly higher in LUAD tissues than in non-tumor
   tissues, and was positively correlated with poor prognosis.
   Phenotypically, CHAC1 overexpression enhanced the proliferation,
   migration, invasion, tumor sphere formation, and glycolysis ability of
   LUAD cells, resulting in tumor growth both in vitro and in vivo.
   Mechanistically, through a shotgun mass spectrometry-based proteomic
   approach and high-throughput RNA sequencing, we found that CHAC1 acted
   as a bridge connecting UBA2 and PKM2, enhancing the SUMOylation of
   PKM2. The SUMOylated PKM2 then transferred from the cytoplasm to the
   nucleus, activating the expression of glycolysis-related genes and
   enhancing the Warburg effect. Lastly, E2F Transcription Factor 1
   potently activated CHAC1 transcription by directly binding to the CHAC1
   promoter in LUAD cells. The results of this study implied that CHAC1
   regulates energy metabolism and promotes glycolysis in LUAD
   progression.

   Subject terms: Cancer metabolism, Tumour biomarkers

Introduction

   Lung cancer is a global health concern, with non-small cell lung cancer
   (NSCLC) accounting for 80–85% of cases [[41]1, [42]2]. Lung
   adenocarcinoma (LUAD) is a major subtype of NSCLC. Patients with LUAD
   have a poor 5-year survival rate of <20%, as many patients are already
   at an advanced stage of the disease when diagnosed [[43]1].
   Additionally, LUAD is characterized by complex mechanisms, hindering
   the development of effective treatments and technological advancements.
   In this context, elucidating the molecular mechanisms underlying the
   development and occurrence of LUAD may facilitate the identification of
   novel therapeutic targets and strategies.

   Notably, few studies have focused on glutathione-specific
   gamma-glutamylcyclotransferase 1 (CHAC1) [[44]3, [45]4]. CHAC1 is known
   as a protein-coding and unfolded protein response (UPR) pathway
   regulatory gene, but also a candidate gene for glutathione (GSH)
   catabolism [[46]5]. The function of CHAC1 in tumorigenesis is
   tumor-type-specific. CHAC1 is upregulated in breast cancer [[47]5,
   [48]6] and advanced clear cell renal cell carcinoma (ccRCC) [[49]7] and
   is associated with poor prognosis. Furthermore, CHAC1 promotes cell
   proliferation and metastasis in uveal melanoma and is a potential
   therapeutic target [[50]8]. In contrast, CHAC1 functions as a tumor
   suppressor gene in primary liver cancer, gastric cancer, prostate
   cancer [[51]9], and human oral squamous cell carcinoma. The
   anti-proliferative effect of metformin in gastric cancer may be partly
   induced by the upregulation of CHAC1 caused via the silencing of lncRNA
   Loc100506691 [[52]10]. The regulation of CHAC1 and subsequent impact on
   tumor growth in hepatocellular carcinoma (HCC) remains debated. It has
   been suggested that CHAC1 is regulated by dihydroartemisinin (DHA),
   which affects its promoter activity and induces ferroptosis in HCC
   cells [[53]11]. Alternatively, CHAC1 may be regulated by MIA3, which
   degrades GSH and leads to HCC progression [[54]12]. However, the
   expression of CHAC1 at the transcriptional and translational levels and
   its correlation with the proliferation and metastasis of LUAD remain
   unclear.

   Metabolic alterations have recently received increasing attention in
   cancer research. Tumor cells exhibit abnormal metabolism,
   preferentially relying on glycolysis even with adequate oxygen, with
   elevated rates of glucose uptake and lactate production, rapid
   production of adenosine triphosphate (ATP) and biomolecules, and
   accompanied by a reduced rate of oxidative phosphorylation. This
   metabolic reprogramming phenomenon, known as the Warburg effect or
   aerobic glycolysis, is a hallmark of cancer cells [[55]13, [56]14]. In
   addition to providing cellular energy, aerobic glycolysis can promote
   uncontrolled proliferation and metastasis of localized tumors [[57]13,
   [58]15], regulate the tumor microenvironment, and promote angiogenesis
   [[59]16] and immune escape [[60]17] to trigger cancer progression.
   Considering that changes in redox status can affect cellular metabolic
   homeostasis, CHAC1, as a glutathione metabolism-related enzyme and an
   upregulated protein in the UPR/endoplasmic reticulum stress process,
   may influence redox status and regulate cellular metabolism [[61]18].
   However, its specific mechanisms remain unclear.

   The glycolytic enzyme pyruvate kinase (PK) is a key rate-limiting
   enzyme that catalyzes pyruvate production. PKM2 contains exon 10 and is
   mainly present in fetal and tumor tissues, promoting aerobic glycolysis
   [[62]19]. PKM2 exists in three forms: inactive monomers, less active
   dimers, and an active tetramer [[63]20]. The tetrameric PKM2 is
   involved in oxidative phosphorylation and rapidly produces ATP. Dimeric
   PKM2, in addition to its role as a glycolysis-related enzyme,
   translocates to the nucleus and acts as a protein kinase and
   co-transcription factor, activating metabolic and proliferative genes
   and displaying non-glycolytic functions [[64]20, [65]21]. However, the
   dynamic changes in PKM2 during tumorigenesis and the molecular
   mechanism of nuclear localization of PKM2 are not fully understood.

   Here, we demonstrated that CHAC1 promotes the nuclear translocation of
   PKM2 by binding to ubiquitin-like modifier activating enzyme 2 (UBA2)
   to activate the small ubiquitin-related modifier modification
   (SUMOylation) of PKM2, which in turn activates glycolytic gene
   expression and promotes LUAD progression. Our study highlighted the
   significance of CHAC1 in LUAD progression and provides new insights
   into carcinogenesis and potential targets for cancer therapy.

Materials and methods

Patients and specimens

   A total of 30 clinical samples (cohort 1) were collected from the
   Fujian Cancer Hospital between January 1, 2018, and October 1, 2022.
   Matched tumor and adjacent normal tissue samples were collected from
   the 30 enrolled patients; these were cut into 1 cm^3 pieces within
   30 min of harvesting, frozen overnight in liquid nitrogen, and stored
   at −80 °C for quantitative real-time PCR and western blot detection.
   LUAD was diagnosed via histopathology. The patients did not receive
   radiotherapy, chemotherapy, targeted therapy, or any other biological
   therapy before surgery.

   The tumor node metastasis (TNM) stage of LUAD was defined according to
   the 7th edition of the American Joint Committee on the Cancer TNM
   staging system. Informed consent was obtained from each patient before
   conducting the study. This study was approved by the Ethics Committee
   of Fujian Cancer Hospital (K2023-139-01).

   A tissue microarray (TMA) (cohort 2) composed of LUAD specimens
   (n = 98) and adjacent normal tissue specimens (n = 82), containing
   detailed clinicopathological information from 81 cases, was purchased
   from Outdo Biotech Ltd. (HLugA180Su08, Shanghai, China) for
   immunohistochemical (IHC) staining. Based on the average CHAC1 IHC
   score (mean value = 5.93), the patients were divided into high (n = 46)
   and low (n = 35) CHAC1 expression groups for clinical analysis. Ethical
   approval for the TMA study was granted by Outdo Biotech’s Clinical
   Research Ethics Committee (SHYJS- CP-1904014).

Analysis of public clinical datasets

   The fragments per kilobase of transcript per million mapped reads of
   the LUAD transcriptome data, and the clinical information for 535 LUAD
   samples and 59 adjacent normal tissues, were obtained from The Cancer
   Genome Atlas (TCGA) database ([66]https://portal.gdc.cancer.gov/),
   including 58 paired samples. ROC analysis was performed using the pROC
   (1.18.0) package, and the results were visualized using the ggplot2
   package in R. Pathway-enrichment analysis was performed using
   Gene-Set-Enrichment Analysis (GSEA) 2.0.9 software
   ([67]http://www.broadinstitute.org/gsea/). The NCBI Gene Expression
   Omnibus datasets [68]GSE31210 and [69]GSE1007 were used to analyze
   CHAC1 expression in cancer and adjacent normal tissues. The
   [70]GSE31210 dataset was used for patient overall survival (OS)
   analysis ([71]http://www.ncbi.nlm.nih.gov/geo/). The University of
   ALabama at Birmingham CANcer data analysis Portal (UALCAN) database
   [[72]22, [73]23] was used to perform correlations between CHAC1 protein
   expression levels and clinical factors, such as the TNM and
   pathological stage, and OS analysis.

Cell transfection

   The human CHAC1 overexpression lentivirus was constructed using the
   pLVZ vector (Zolgene, Fuzhou, China). Lentiviral shRNAs targeting CHAC1
   were constructed using GV344 and pLV vectors (GeneChem, Shanghai,
   China), lentiviral shRNAs targeting PKM2 were constructed using GV344
   and pLV vectors (Hanbio, Shanghai, China), respectively. To stabilize
   transfection, LUAD cells were seeded in 6-well plates at ~60%
   confluence overnight and then infected with the corresponding
   lentivirus in the presence of polybrene (4 µg/ml). Twenty-four hours
   after transfection, the medium was replaced with fresh medium.
   Transfected cells were selected with puromycin (5 µg/ml). siRNAs
   targeting E2F1, MAZ, and PKM2 were synthesized by HippoBio Technology
   (Huzhou, China). siRNAs targeting UBA2 were synthesized by Hanbio. The
   HA-tagged PKM2 wild-type plasmid and the His-tagged PKM2^I267/268A
   plasmid were cloned into the GV658 vector (GeneChem). The Flag-tagged
   SUMO1 plasmid was cloned into the CMV vector (Hanbio). The 2000-bp
   sequence upstream of the CHAC1 promoter was inserted into the pgl4.10
   vector to generate a luciferase reporter plasmid. Transient
   transfection was performed using Lipofectamine 3000 (Invitrogen,
   Carlsbad, CA, USA) according to the manufacturer’s protocol. Cells were
   harvested 48–72 h after transfection. The sequences of shRNAs and
   siRNAs used in this study are listed in Supplementary Table [74]S1.

Immunohistochemical (IHC) staining

   IHC staining was performed for tissue sections, TMA, and xenograft
   tumors. Briefly, sections were incubated with primary antibodies to
   CHAC1, PKM2, and Ki67 overnight at 4 °C. Normal goat serum was used as
   a negative control. Sections were washed and incubated with a
   biotinylated anti-rabbit secondary antibody (Roche, Switzerland),
   followed by streptavidin-horseradish peroxidase complex (Dako,
   Carpinteria, CA, USA). The slices were soaked in 3,3-diaminobenzidine,
   counterstained with 10% Mayer’s hematoxylin, dehydrated, and mounted.
   TMA images were captured at 40× magnification using an Aperio GT450
   scanner (Leica Biosystems). Sliced xenograft tumor images were captured
   at 100× magnification using the same scanner. The staining intensity
   was scored as follows: no intensity, 0; weak intensity, 1+; moderate
   intensity, 2+; and strong intensity, 3+. The staining extent score,
   based on the percentage of positively stained cells, was categorized as
   0 (0%), 1 (1–25%), 2 (26–50%), 3 (50–75%), or 4 (76–100%). The final
   histological score (H-score) was calculated by multiplying the staining
   intensity score with the staining extent score. These scores were
   determined independently by two senior pathologists who were blinded to
   the clinical and pathological information. The primary antibodies used
   are listed in Supplementary Table [75]S2.

Tumor sphere formation

   Tumor spheroid formation was performed as previously described
   [[76]24]. Cells (4 × 10^4) were seeded in 6-well ultra-low adsorption
   plates and then cultured in a saturated humidity atmosphere of 5% CO[2]
   at 37 °C to form tumor spheres. After 2 weeks, images of the cells were
   scanned using an inverted microscope at 200× magnification. The number
   of tumor spheres was counted and plotted, and the percentage of tumor
   spheres >100 µm in diameter was calculated using Image J software.

Wound-healing assay

   For in vitro wound-healing assays, transfected A549 and PC9 cells were
   seeded into 6-well plates. The indicated cells were grown to confluency
   as monolayers and then scratched with a 10-µl pipette tip. The cells
   were washed with PBS to remove floating cells, and the scratch was
   imaged and recorded at 0 h. The cells were then incubated in 1% serum
   medium for 48 h, imaged again, and the percentage of the scratched area
   covered with cells was determined using Image J software.

Migration and invasion assays

   Migration and invasion assays were performed using transwell chambers
   with 8.0-μM pore sizes (Corning, USA). For the migration assays,
   1 × 10^5 transfected cells were resuspended in 200 µl serum-free 1640
   medium and added to the upper chambers, while 2 × 10^5 transfected
   cells in chambers coated with Matrigel were used for the invasion
   assay. For both assays, the bottom chambers were filled with RPMI 1640
   medium containing 10% FBS. After 24 h, the migrating and invasive cells
   on the membranes were stained with 0.5% crystal violet and photographed
   under a microscope, and the cell numbers were counted using Image J
   software.

Flow cytometry (FACS) for cell-cycle analyses

   Cell-cycle analysis was performed using the Cell-Cycle and Apoptosis
   Analysis Kit (Meilunbio, China, MA0334). A total of 1 × 10^6 cells were
   collected, and 1 ml PBS was added to induce the resuspension of single
   cells. Precooled absolute ethanol (3 ml) was slowly added, and the
   mixture was allowed to stand overnight at 4 °C, after which it was
   centrifuged at 1000 × g for 3–5 min. The mixture absorbed the
   supernatant, and 1 ml PBS was again added to cause resuspension.
   Staining buffer: PI (20×): RNase A (50×) = 100:5:2 was used to prepare
   the working solution for staining. Staining working solution (500 μl)
   was added to each tube of the cell sample, and red fluorescence was
   detected at an excitation wavelength of 488 nm via flow cytometry after
   a warm bath at 37 °C in the dark for 30 min.

Immunoprecipitation (IP) analysis

   The cells were lysed with RIPA lysis buffer (Beyotime, Shanghai, China)
   and supplemented with a protease inhibitor cocktail (Beyotime). Cell
   lysates were incubated with specific primary antibodies and protein A/G
   magnetic beads (Smart-Lifesciences, China, SM015005), and normal IgG
   served as the control. The magnetic beads were washed thrice with TBS
   buffer (50 mM Tris-HCl and 150 mM NaCl, pH 7.4), boiled, and analyzed
   by using western blotting.

Silver staining and shotgun mass spectrometry (MS)

   IP was performed as previously described. Bound proteins were separated
   using SDS–PAGE (LABLEAD, Beijing, China) and stained using a Silver
   Staining Kit (Beyotime, Shanghai, China). The protein solution or
   strips were digested into peptide mixtures via protease and then
   separated by high-performance liquid chromatography (HPLC), followed by
   serial transfer into an Orrap Discovery mass spectrometer (Thermo) for
   LC–MS/MS analysis. After the peptides were ionized in the mass
   spectrometer, the mass-charge ratio (m/z) of each peptide was obtained
   by detector analysis, and a mass spectrometer further bombarded the
   peptide ions to obtain the secondary mass spectrum signal. Mascot2.6,
   Proteome Discoverer2.2, NCBInr, and UniProt were used for protein
   identification.

High-throughput RNA sequencing

   Total RNA was extracted using the TRIzol reagent (Invitrogen) according
   to the manufacturer’s instructions. RNA integrity and concentration
   were assessed using the RNA Nano 2100 assay kit (Agilent Technologies,
   USA) and Bioanalyzer 6000 system. Total RNA was >2 μg/sample. Libraries
   were constructed using Illumina’s NEBNext® UltraTM RNA Library Prep
   Kit. The libraries were quantified and detected using a Qubit2.0
   Fluorometer with an Agilent 2100 bioanalyzer. Illumina sequencing was
   performed after a qualified library inspection. The sequencing
   sequences were converted into sequence data (reads) using CASAVA base
   recognition, and the files were in FASTQ format. The HISAT tool was
   used to align sequences to the reference genome. StringTie (1.3.3b)
   [[77]25] was used for novel gene prediction. FeatureCounts (1.5.0-p3)
   were used to calculate the reads mapped to each gene, and DESeq2
   software (version 1.16.1) was used for differential expression scores.
   The R package “edgeR” was utilized to explore the differentially
   expressed genes with the fold change criteria >2 and p value < 0.05.
   ClusterProfiler (version 3.4.4) software was used to perform GO and
   KEGG pathway-enrichment analyses.

Immunofluorescence staining

   The cells were seeded in a 24-well chamber slide. The next day, the
   cells were fixed with 4% formaldehyde, permeabilized with 0.3% Trion
   X-100, blocked with QuickBlock™ blocking solution (Beyotime, China),
   and incubated with the specific antibodies overnight. After washing
   with PBS, the cells were incubated with Alexa Fluor 488 and/or Alexa
   Fluor 594 (1:1000 dilution; Abcam, UK) for 2 h. The nuclei were stained
   with DAPI, and the cells were imaged using a laser confocal microscope
   (NIKON Eclipse Ti, Japan). The primary antibodies used are listed in
   Supplementary Table [78]S2.

Chromatin immunoprecipitation (ChIP) assays

   A ChIP kit (Thermo Fisher Scientific, USA) was used according to the
   manufacturer’s protocol. Briefly, cells (1 × 10^7 cells/condition) were
   crosslinked, lysed, and sonicated to obtain DNA samples. DNA fragments
   were immunoprecipitated with an E2F1 antibody (Proteintech, China), and
   the binding of E2F1 to the CHAC1 promoter region was analyzed via
   RT-qPCR using specific primers for ChIP-enriched DNA fragments and
   normalized to control IgG (Supplementary Table [79]S1).

Dual-luciferase reporter assay

   Cells were seeded in 96-well plates and co-transfected with plasmids
   containing the E2F1 and MAZ sequences cloned into pCMV, pGL4.10
   luciferase reporter vector containing the CHAC1 promoter (Promega,
   USA), and pRL-TK. Lysates were prepared 48 h after transfection, and
   the luciferase activity was measured using a dual-luciferase reporter
   assay system (Promega, USA) according to the manufacturer’s protocol.
   The luciferase activity was calculated as the ratio of firefly to
   Renilla luciferase activity.

Glycolysis analysis

   Assay kits for glucose uptake colorimetry, lactate colorimetry, ATP,
   hexokinase (HK), fructose phosphokinase (PFKP), PK, and LDH activity
   (Jiancheng Corporation, Nanjing, China) were used according to the
   manufacturer’s instructions.

   For the glucose-uptake experiment and lactic acid colorimetric assay
   experiment, cells were collected at a density of more than
   1 × 10^6 cells/ml and washed with isotonic buffer (0.1 mol/L, pH 7–7.4
   phosphate buffer). The supernatant was discarded, and the cell
   precipitate was retained. Following this, 0.2–0.3 ml of homogenization
   medium (0.1 mol/L, pH 7–7.4 phosphate buffer) was added for
   homogenization. The prepared homogenate was added to a 96-well plate,
   and the absorbance at 505 nm was measured using a microplate reader. In
   the ATP assay experiment, 1 × 10^6 cells were collected, 300–500 µl hot
   double-distilled water was added, and the cells were homogenized and
   broken in a hot water bath (90–100 °C). The water bath was boiled for
   10 min. The absorbance at a wavelength of 636 nm was measured by using
   a microplate reader. In the HK kit, fructose PFKP-activity assay, PK
   test and LDH activity test experiments, 5 × 10^6 cells were collected
   and added to 1 ml of the prepared reagent and lysed via ultrasound, the
   supernatant was collected, and the absorbance was determined using a
   microplate reader.

Seahorse-XF24-extracellular-acidification rate (ECAR) and oxygen consumption
rate (OCR)

   ECAR and OCAR measurements were performed using a hippocampal XF24
   extracellular flux analyzer (Hippocampal, USA). After trypsinization of
   the cells, cancer cells were seeded in hippocampal XF24 cell plates
   (1 × 10^4/well) and cultured for 24 h in a humidified 37 °C, 5% CO[2]
   incubator. The cell plates were then incubated at 37 °C in a 0% CO[2]
   incubator for 60 min. For ECAR measurements, the Seahorse XF Glycolytic
   Stress Test Kit was used, with the medium replaced by Seahorse XF Base
   Medium sequentially supplemented with glucose (10 mM), oligomycin (1 μM
   and 2 μM), and 2-deoxyglucose (2-DG, 80 mM) according to the
   manufacturer’s protocol. OCR was measured using the Seahorse XF cell
   mito stress test kit, with the medium replaced by Seahorse XF base
   medium sequentially supplemented with oligomycin (1.5 μM), FCCP
   (0.5 μM), and rotenone (0.5 μM) following the manufacturer’s
   instructions.

Mouse xenograft study

   The animal experimental protocol involved in this study was reviewed
   and approved by the Animal Ethics Committee of Fujian Medical
   University (No. IACUC FJMU 2023-0109). BALB/c nude mice (4–6-week-old)
   were obtained from GemPharmatech Co., Ltd., Jiangsu, China. All mice
   were fed a standard rodent diet and maintained under specific
   pathogen-free conditions with an ambient temperature of 23 °C and a
   12-h/12-h light/dark cycle. Nude mice were randomly assigned to groups
   and injected with 5 × 10^6/100 µl A549 cells transfected with
   overexpression lentivirus (n = 6/group), PC9 cells transfected with
   knockdown lentivirus (n = 6/group), and A549 cells transfected with
   CHAC1 overexpression lentivirus combined with PKM2-knockdown lentivirus
   (n = 8/group) to establish a subcutaneous xenograft model. The longest
   (A) and shortest (B) diameters of the grafts were measured weekly, and
   the tumor volume was monitored for 4 weeks after inoculation
   (calculated as 0.5 × A × B^2). After 4 weeks, or when the tumors
   reached a diameter of 1.5 cm, the mice were euthanized, and the
   subcutaneous tumors were harvested. The specimens were weighed and
   subjected to IHC staining.

Statistical analysis

   Statistical analyses were performed using GraphPad Prism 7.0. All the
   data are presented as the mean ± SD and represent at least three
   independent experiments. Spearman’s rank correlation test was used to
   analyze the correlation between gene expression in tissue samples.
   ANOVA was used to compare the means between the treatment groups.
   Student’s t-test was used for unpaired observations, with p < 0.05
   being considered significant. Correlation analyses were performed using
   the R-value test. OS was defined as the time from surgery to death from
   LUAD or the last follow-up. The Kaplan–Meier method and log-rank test
   were used for survival analysis. All the tests were two-tailed.

Results

Upregulation of CHAC1 correlates with aggressive phenotype and poor prognosis
in LUAD

   To identify the potential clinical significance of GSH-catabolism genes
   in human LUAD, we first analyzed the mRNA expression levels of 11
   GSH-catabolism genes in primary tumors and adjacent normal tissues
   using the TCGA dataset. The three most highly expressed genes were
   DPEP1 (log[2]FC = 1.61, p < 0.001), CHAC2 (log[2]FC = 1.20, p < 0.001),
   and CHAC1 (log[2]FC = 1.04, p < 0.001) (Fig. [80]1A and Supplementary
   Table [81]S3). In the 58 matched samples, the expression of CHAC1,
   CHAC2, and DPEP1 was significantly higher in LUAD tissues than in the
   adjacent normal lung tissues (p < 0.001) (Fig. [82]1B). Notably,
   patients with LUAD and high CHAC1 expression had worse OS than those
   with low CHAC1 expression (p < 0.05). However, CHAC2 and DPEP1
   expression did not affect LUAD OS (Fig. [83]1C). In addition, the
   results of CHAC1 in the TCGA dataset were verified in the GEO datasets,
   including [84]GSE31210 and [85]GSE10072 (Fig. [86]1D). Moreover, high
   CHAC1 expression was associated with advanced TNM stage and N stage
   (p < 0.05) (Supplementary Fig. [87]1). We also found that CHAC1
   expression was highly accuracy in predicting LUAD (AUC = 0.793, Fig.
   [88]1E).

Fig. 1. High CHAC1 expression was significantly associated with poor
prognosis in LUAD.

   [89]Fig. 1
   [90]Open in a new tab

   A Transcription levels of GSH catabolic genes in LUAD tissues (n = 535)
   and adjacent normal tissues (n = 59) from TCGA database (screened by
   logFC > 1, p < 0.05). B Transcription levels of DPEP1, CHAC2, and CHAC1
   in 58 matched tumor and adjacent normal tissue samples of LUAD in TCGA
   database. C Kaplan–Meier curves for OS based on high and low DPEP1,
   CHAC2, and CHAC1 levels in TCGA database. D GEO datasets ([91]GSE31210,
   [92]GSE10072) were used to analyze CHAC1 expression and OS in LUAD. E
   Diagnostic ROC was used to analyze CHAC1 expression in predicting LUAD
   accuracy. F, G IHC staining for CHAC1 in LUAD tissue microarray. H
   Kaplan–Meier curves for OS according to high and low CHAC1 levels of
   LUAD tissues in cohort 2. I Protein levels of CHAC1 were detected in
   primary LUAD tissues (T) and corresponding adjacent nontumoral lung (N)
   tissues. J mRNA and protein expression levels of CHAC1 in Beas-2b and
   LUAD cell lines were detected by using RT-qPCR and western blot.
   Beas-2b was used as internal control. Three independent experiments
   were performed, and data are means ± SD from one representative
   experiment (n = 3). Values below western blot bands are optical density
   of signal. *p < 0.05, **p < 0.01, ***p < 0.001.

   Subsequently, we used a TMA containing 98 LUAD tissues and adjacent
   normal lung tissues to further evaluate the expression of CHAC1 in
   LUAD. IHC staining showed that CHAC1 was mainly distributed in the
   cytoplasm and cell membrane of LUAD cells, and its expression was
   higher than that in normal lung tissues (Fig. [93]1F, G). These results
   are consistent with those of the public database analysis. Survival
   analysis revealed that patients with LUAD with high CHAC1 expression
   had significantly worse OS than those with low CHAC1 expression
   (p = 0.017, Fig. [94]1H). Moreover, in 81 samples with clinical
   information, CHAC1 expression was significantly associated with a worse
   T stage (Supplementary Table [95]S4) and was an independent prognostic
   factor for poor prognosis in patients with LUAD (Supplementary Table
   [96]S5).

   Next, the CHAC1 mRNA and protein expression levels in vitro were
   assessed. Thirty fresh tissue specimens were collected, and CHAC1 was
   significantly more highly expressed in cancerous tissues than in
   paracancerous tissues (p < 0.001, Fig. [97]1I). Furthermore, the
   expression of CHAC1 in the LUAD cell lines was higher than that in
   normal lung epithelial cells (Beas-2b) (Fig. [98]1J). These results
   indicate a positive correlation between CHAC1 expression and the
   aggressive progression and poor prognosis of LUAD.

CHAC1 promotes LUAD proliferation and metastasis

   To explore the functional role of CHAC1 in LUAD, we constructed a
   CHAC1-FLAG structure by fusing the tag with a full-length CHAC1
   transcript. Stable CHAC1-overexpressing cell lines (OE) were generated
   from the A549 cell line, which exhibits relatively low CHAC1
   expression, with Vector used as the positive control (VE).
   CHAC1-knockdown cell lines (sh1, sh2, and sh3-CHAC1) were established
   from the PC9 cell line, which has relatively high CHAC1 expression, and
   a negative control (NC) was prepared. The overexpression and knockdown
   efficiency of CHAC1 in these cell sublines was detected via western
   blotting and RT-qPCR; sh2- and sh3-CHAC1 showed greater inhibition of
   CHAC1 expression (Fig. [99]2A and Supplementary Fig. [100]2A) and were
   renamed shCHAC1 #1 and shCHAC1 #2 for subsequent experiments. As shown
   in Fig. [101]2B, CHAC1 was mainly distributed in the cell membrane and
   cytoplasm.

Fig. 2. CHAC1 promotes proliferation and metastasis of LUAD cells.

   [102]Fig. 2
   [103]Open in a new tab

   A CHAC1 overexpression in A549 and shCHAC1 PC9 cells was constructed
   via transfecting lentivirus with overexpression of shRNA and verified
   by using western blotting and qRT-PCR. B-actin was used as a loading
   control. B IF staining was used to detect the distribution of CHAC1 in
   cells. Scale bars, 10 µm. C Colony-formation assay, performed to detect
   the effect of overexpression or knockdown of CHAC1 on cell
   proliferation. D Flow cytometry was used to analyze the effect of
   overexpression or knockdown of CHAC1 on cell-cycle progression. E–H
   Wound-healing assay and transwell assay were used to detect cell
   migration and invasion abilities after overexpression or knockdown of
   CHAC1. Scale bars, 200 µm. I Western blotting was used to detect the
   effect of overexpression or knockdown of CHAC1 on EMT. Results are
   shown as means ± SD. Values below western blot bands are optical
   density of signal. *p < 0.05, **p < 0.01, ***p < 0.001, ns no
   statistical significance.

   Based on the clone-formation assay, CHAC1 overexpression significantly
   promoted LUAD cell proliferation, whereas CHAC1 depletion markedly
   decreased the proliferative capacity of LUAD cells (Fig. [104]2C and
   Supplementary Fig. [105]2B, C). Pathway-enrichment analysis showed that
   the cell cycle G1_S phase transition (p < 0.05) and G2_M transition
   (p < 0.05) were significantly positively correlated with the expression
   level of CHAC1 in the TCGA dataset (Supplementary Fig. [106]2D). Thus,
   we further performed cell-cycle analysis and found that the proportion
   of cells in the S and G2/M phases increased in CHAC1-overexpressing
   A549 cells, whereas the CHAC1-knockdown group exhibited a larger
   proportion of cells in the G0/G1 phase (Fig. [107]2D). These results
   suggest that CHAC1 is an important regulator of cell proliferation in
   LUAD.

   In addition, the migration and invasion capacities of LUAD cells were
   significantly enhanced when CHAC1 was overexpressed, with opposite
   trends being observed under CHAC1 depletion (Fig. [108]2E–H and
   Supplementary Fig. [109]2E, F). Since the epithelial–mesenchymal
   transition (EMT) process is integral to tumor metastasis [[110]26], we
   further investigated the relationship between CHAC1 expression and EMT.
   As shown in Fig. [111]2I, CHAC1 overexpression inhibited E-cadherin
   expression and enhanced the expression levels of N-cadherin, MMP2, and
   MMP9, while the knockdown of CHAC1 had the opposite effect. We also
   detected the effect of CHAC1 expression on the tumor sphere-forming
   capacity. Compared to the control group, the overexpression of CHAC1 in
   A549 cells increased the size and number of tumor spheres. Conversely,
   the depletion of CHAC1 in PC9 cells reduced the size and number of
   tumor spheres (Supplementary Fig. [112]2G). Collectively, CHAC1
   expression contributes to the proliferation and metastasis of LUAD
   cells.

CHAC1 enhances glycolysis of LUAD cells

   To investigate the molecular mechanism of CHAC1 in LUAD, we performed
   an unbiased transcriptomic analysis using RNA-seq in A549 cells with or
   without overexpression of CHAC1. The results showed that the expression
   levels of 768 genes were affected by CHAC1 overexpression, of which 310
   genes were upregulated and 458 genes were downregulated (Fig. [113]3A).
   The KEGG and GO analyses revealed that the pathways were enriched in
   these differentially expressed genes (Supplementary Fig. [114]3A). The
   GSEA revealed that the target genes were significantly involved in the
   energy-metabolism signaling pathway (Fig. [115]3B). GSEA was also
   performed on the mRNA expression profile of LUAD in the TCGA dataset,
   and showed that the activity of the energy metabolism-related pathways,
   including glycolysis/gluconeogenesis and the citrate cycle (TCA cycle),
   was positively correlated with CHAC1 expression (NES > 1.5, p < 0.05,
   Fig. [116]3C).

Fig. 3. CHAC1 is essential for the glycolytic activity of LUAD.

   [117]Fig. 3
   [118]Open in a new tab

   A Volcano plots depict upregulated (red) and downregulated (blue) genes
   in CHAC1 OE cells. B Differential genes in CHAC1 OE cells were
   subjected to GSEA. C The pathways involved in CHAC1 expression were
   analyzed based on TCGA database. D Glucose uptake, lactate production,
   and ATP levels in CHAC1, OE, and knockdown cells were quantified. E
   Enzymatic activity reactions of HK, PFK, and PK in CHAC1, OE, and
   knockdown cells were detected. F Relative HK, PFKP, and PKM2 mRNA
   levels in CHAC1, OE and knockdown cells were analyzed. G ECAR of
   overexpression and knockdown of CHAC1 were monitored, and glycolysis,
   glycolytic capacity, and glycolytic reverse were calculated. Results
   are shown as means ± SD. NES normalized enrichment score, FDR false
   discovery rate. *p < 0.05, **p < 0.01, ***p < 0.001, ns no statistical
   significance.

   Next, we examined the association between CHAC1 expression and
   glycolytic phenotype. The upregulation of CHAC1 enhanced glucose
   uptake, lactate production, and ATP levels, whereas the downregulation
   of CHAC1 showed the opposite results (Fig. [119]3D). Since the
   glycolytic pathway is controlled via several key glycolytic
   rate-limiting enzymes, including HK, PFK, and PK [[120]27], we aimed to
   determine whether CHAC1 expression affects key glycolytic rate-limiting
   enzymes. The results showed that the downregulation of CHAC1 enhanced
   PK activity, while the knockdown of CHAC1 had the opposite effect;
   however, the activities of HK and PFK were not affected by the
   expression of CHAC1 (Fig. [121]3E). Interestingly, we detected the
   glycolytic rate-determining enzymes HK2, PFKP, and PKM2 via
   quantitative PCR. The results showed that interfering with the
   expression of CHAC1 did not affect the mRNA levels of HK2, PFKP, or
   PKM2 (Fig. [122]3F). Moreover, we detected the impact of CHAC1
   expression on ECAR, an indicator of overall glycolytic flux and OCAR,
   an indicator of mitochondrial respiration. The upregulation of CHAC1
   enhances glycolysis and glycolytic capacity in LUAD cells, evidenced by
   a significant increase in ECAR. Conversely, CHAC1 knockdown produces
   the opposite effect. Notably, alterations in CHAC1 expression do not
   significantly impact OCAR in LUAD cells (Fig. [123]3G and Supplementary
   Fig. [124]3B). These results indicated that CHAC1 may control
   glycolytic metabolism by specifically regulating PK activity.

   CHAC1 is also known to act as a γ-glutamylcyclotransferase, leading to
   GSH depletion and possibly inducing ferroptosis [[125]25]. Therefore,
   we examined whether altering CHAC1 expression in LUAD cells would
   result in GSH depletion and ferroptosis. Results showed that CHAC1
   overexpression reduced the total intracellular GSH and induced
   intracellular lipid peroxidation, while CHAC1 deficiency had the
   opposite effect (Supplementary Fig. [126]3C, D). However, we examined
   the proportion of 7-AAD-positive cells and found that alterations in
   CHAC1 expression did not significantly affect cell death (Supplementary
   Fig. [127]3E). Moreover, the additional use of ferroptosis inhibitor
   (ferrostatin-1, Fer-1) could not abolish the enhanced cellular
   glycolytic capacity caused by CHAC1 overexpression (Supplementary Fig.
   [128]3F), which mean that the enhancement of glycolytic pathway is not
   dependent through the glutathione depletion role of CHAC1.

CHAC1 interacts with PKM2 and prevents its degradation

   Since we have previously provided evidence that CHAC1 increases the
   glycolytic capacity of LUAD cells, we next investigated the specific
   mechanism. Silver staining and shotgun MS were performed to detect
   CHAC1-interacting molecules (Fig. [129]4A). The isolation of
   CHAC1-interacting protein complexes from CHAC1 OE A549 cells took place
   using anti-FLAG monoclonal antibodies and paramagnetic beads. Affinity
   purification and MS analysis detected 400 binding proteins with CHAC1,
   with GSEA being mainly enriched in pathways including metabolic
   processes, glycolytic processes, and splicing-related proteins
   (Supplementary Tables [130]S6 and [131]S7). We intersected 31
   differential expressed proteins in the “glucose metabolism” pathway
   from RNA-seq with 14 CHAC1-binding proteins in the MS “glycolytic
   process” and obtained 10 CHAC1- and glycolysis-related proteins (Fig.
   [132]4B).

Fig. 4. CHAC1 interacts with and stabilizes PKM2.

   [133]Fig. 4
   [134]Open in a new tab

   A Identification of CHAC1-interacting proteins by using silver
   staining. Indicated bands were sequenced via mass spectrometry
   analysis. B Proteins in LC–MS “glycolytic process” pathway intersect
   with those in “glucose metabolism” pathway in RNA-seq. C Endogenous
   interaction between CHAC1 and PKM2 in LUAD cells was detected via
   immunoprecipitation (IP). D Co-IP was conducted in LUAD cells
   transfected with Flag-CHAC1 lentivirus. E IP was conducted in LUAD
   cells transfected with HA-PKM2 lentivirus. F CHAC1 and PKM2
   co-localization were determined via confocal microscopy scanning of IF
   staining in LUAD cells. DAPI was used for nuclear staining. Scale
   bar = 10 µm. G Protein levels of PKM2 in overexpression or knockdown of
   CHAC1 LUAD cells. H Protein levels of CHAC1 in overexpression or
   knockdown of PKM2 LUAD cells. Values below western blot bands are
   optical density of signal.

   GAPDH and PKM exhibited the highest MS coverage (Supplementary Table
   [135]S6), and the spectrum of PKM is shown in Supplementary Fig.
   [136]4. Since PKM2 is an essential rate-limiting metabolic enzyme in
   glycolysis [[137]28], we focused on PKM2 for further experiments. The
   results showed that endogenous PKM2 bound to CHAC1 (Fig. [138]4C). We
   transfected an exogenous FLAG-tagged CHAC1 lentivirus and found that
   CHAC1-FLAG could bind to PKM2 (Fig. [139]4D). Next, we expressed
   HA-tagged PKM2 in PC9 and A549 cells and found that FLAG-tagged CHAC1
   bound to HA-tagged PKM2 (Fig. [140]4E). IF analysis revealed that CHAC1
   interacted with PKM2 in both the cytoplasm and the nucleus (Fig.
   [141]4F).

   Given that CHAC1 expression affects PK activity (Fig. [142]3E), the
   question is whether this is achieved by affecting PKM2 protein
   expression. Therefore, we examined PKM2 protein expression by
   interfering with CHAC1 expression. The results showed that
   overexpressing or knocking down CHAC1 did not affect PKM2 protein
   expression (Fig. [143]4G). Similarly, interfering with PKM2 expression
   did not affect CHAC1 protein expression (Fig. [144]4H). These results
   showed an interaction between CHAC1 and PKM2, which did not change the
   expression of PKM2 but affected PK activity.

CHAC1 induces SUMOylation of PKM2 through UBA2/SUMO1 axis

   It has been suggested that PKM2 may affect its activity by interfering
   with its subcellular localization by SUMOylation [[145]29] and SUMO1
   promotes the glycolytic ability of tumor cells [[146]30]. Therefore, we
   explored whether CHAC1 mediates the SUMOylation of PKM2 and further
   regulates glycolytic metabolism. First, SUMO1 protein was extracted
   from A549 and PC9 cells for immunoprecipitation and analyzed by liquid
   chromatography–tandem mass spectrometry (LC–MS/MS). LC–MS/MS analysis
   revealed that PKM was one of the SUMOylated proteins (Supplementary
   Fig. [147]5A). Importantly, endogenous PKM2 SUMOylation in LUAD cells
   was detected by co-immunoprecipitation (IP). PKM2 SUMOylation was
   similarly detected when PKM2 (hemagglutinin [HA]) and SUMO1 (Flag) were
   overexpressed (Fig. [148]5A, B).

Fig. 5. CHAC1 induces SUMOylation of PKM2 protein through UBA2/SUMO1 axis.

   [149]Fig. 5
   [150]Open in a new tab

   A Endogenous interaction between PKM2 and SUMO1 in LUAD cells was
   detected via IP. B Exogenous interaction between HA-PKM2 and FLAG-SUMO1
   in LUAD cells was detected via IP. C SUMOylation of PKM2 in CHAC1
   overexpression cells was detected via IP. D SUMOylation of PKM2 in
   shCHAC1 cells was detected via IP. E SUMO inhibitor GA (50 µm) was
   added to detect the effect of SUMOylation on PKM2 protein. F
   Intersection of GSEA “GOBP_PROTEIN_SUMOYLATION” signature and LC–MS was
   used to identify CHAC1-binding proteins affecting PKM2 SUMOylation. G
   IP was used to detect interactions between CHAC1, UBA2, and PKM2. H The
   SUMOylation of PKM2 was detected by knocking down UBA2. I The
   interaction between CHAC1, PKM2, and UBA2 was examined by knocking down
   CHAC1. J The interaction between CHAC1, UBA2, and PKM2 was detected by
   knockdown of CHAC1 combined with overexpression of UBA2. IB immunoblot,
   IP immunoprecipitation.

   Subsequently, we found that overexpression of CHAC1 resulted in a
   distinct shift band of PKM2 at 95 kDa. Knockdown of CHAC1 reduced the
   intensity of the 95 kDa shift band (Fig. [151]5C, D). SUMOylation of
   PKM2 resulted in a shift band at 95 kDa [[152]31]. Therefore, to
   demonstrate that PKM2 undergoes SUMOylation, we added Ginkgolic acid
   (GA), a small molecule that inhibits SUMOylation by targeting the
   SUMO-activating enzyme E1 [[153]32], while overexpressing CHAC1. As
   shown in Fig. [154]5E, GA inhibited the distinct shift band of PKM2 at
   95 kDa caused by CHAC1 overexpression, further supporting enhanced PKM2
   SUMOylation. These results indicate that CHAC1 promotes the binding of
   PKM2 to the SUMO1 protein and its SUMOylation.

   Next, we explored whether SUMOylation-related proteins were presented
   in the LC–MS analysis. The intersection of LC–MS and the
   “GOBP_PROTEIN_SUMOYLATION” signature in GSEA was taken, and four
   potential CHAC1-binding proteins that affected PKM2 SUMOylation were
   identified (Fig. [155]5F). UBA2 is a component of SUMO-activating
   enzyme 1 (SAE1), which regulates the critical activation of SUMO
   proteins. IP experiments revealed an interaction between FLAG-CHAC1,
   UBA2, and PKM2 (Fig. [156]5G). Therefore, we speculated that CHAC1
   interacts with UBA2 to promote the SUMOylation of PKM2. IF staining
   showed that CHAC1 co-localized with UBA2 in the cytoplasm
   (Supplementary Fig. [157]5B). When we knocked down UBA2, the SUMOlytion
   level of PKM2 decreased (Fig. [158]5H and Supplementary Fig. [159]5C).
   Then we further examined their interactions, the interaction between
   UBA2 and PKM2 was significantly attenuated by CHAC1 knockdown (Fig.
   [160]5I). The interaction between UBA2 and PKM2 was not restored by
   UBA2 overexpression CHAC1-knockdown cells (Fig. [161]5J). These results
   suggest that CHAC1 may act as a bridge connecting UBA2 and PKM2,
   thereby facilitating their interaction and enhancing PKM2 SUMOylation.

CHAC1 regulates translocation of PKM2 into nucleus via SUMOylation

   SUMO-interacting motif (SIM) is a SUMOylation recognition motif of
   target proteins. There are three potential SIM sites in PKM2, among
   which IKII265-268 is mutated to PKM2^I267&268A (MT), eliminating the
   SUMO1-induced SUMOylation of PKM [[162]33], which affects PKM2 dimer
   formation and nuclear translocation [[163]29]. Hence, we hypothesized
   that CHAC1 regulates PKM2 SUMOylation and nuclear translocation and
   that PKM2 acts as both a protein kinase and a co-transcription factor
   in the nucleus. We examined the localization of PKM2 in
   CHAC1-overexpression cells. Nucleocytoplasmic separation experiments
   showed that CHAC1 overexpression enhanced the expression level of PKM2
   in the nucleus but did not change the overall PKM2 protein level (Fig.
   [164]6A). Similar results were obtained with the IF assay (Fig.
   [165]6B). When transfected with the PKM2 WT plasmid, PKM2 nuclear
   translocation was enhanced, whereas transfection with the PKM2 MT
   plasmid abolished this effect (Fig. [166]6C, D). These results indicate
   that CHAC1 modifies specific sites through SUMOylation to promote the
   nuclear translocation of PKM2.

Fig. 6. CHAC1 regulates the nuclear translocation of PKM2 via SUMOylation.

   [167]Fig. 6
   [168]Open in a new tab

   A Nuclear and cytosolic protein lysates prepared from A549 cells stably
   expressing mock or CHAC1 were assayed via western blotting. B
   Subcellular localization of PKM2 in LUAD cells that stably expressed
   mock or CHAC1 OE was observed via IF. C Cells were transfected with
   HA-tagged PKM2 (WT), His-tagged PKM2^I267&268A (MT), and PKM2
   localization was detected by using western blot. D Subcellular
   localization of PKM2 in LUAD cells transfected with PKM2 WT and PKM2 MT
   was observed via IF. E, F, M PKM2 WT or PKM2 MT were transfected into
   LUAD cells simultaneously overexpressing CHAC1 and siPKM2,
   respectively. E Colony-formation assay was used to detect cell
   proliferation. F Relative glucose uptake, lactate production, ATP level
   were detected. G, H RT-qPCR and western blot were used to detect
   relative mRNA and protein levels of GLUT1, ENO1, and LDHA in CHAC1 OE
   and knockdown cells. I The correlation between CHAC1 expression and the
   expression of GLUT1, ENO1, and LDHA in 30 LUAD tissue samples. J Detect
   LDH activity in LUAD cells using a lactate dehydrogenase assay kit. K
   Western blot was used to detect phosphorylation of STAT3^Y705 and
   histone H3^T11. L The phosphorylation level of STAT3 was detected after
   the addition of 7a. M Western blot was used to detect phosphorylation
   of STAT3^Y705 and histone H3^T11. Results are shown as means ± SD.
   Values below western blot bands are optical density of signal.
   **p < 0.01, ***p < 0.001.

   To elucidate whether the progression induced by CHAC1 depends on the
   nuclear translocation of PKM2. We stably transfected a
   CHAC1-overexpressing lentivirus in cells with PKM2 knockdown. Results
   showed that PKM2 knockdown significantly reduced cell proliferation and
   migration, and CHAC1 overexpression could not reverse these effects
   (Supplementary Fig. [169]S6A, B). These results suggest that PKM2 is a
   key downstream effector of CHAC1 during the regulation of cell
   proliferation and metastasis.

   Then, we transfected siPKM2, PKM2 WT, and PKM2 MT plasmids into cells
   stably overexpressing CHAC1. The results showed that PKM2 inhibited the
   enhanced cell proliferation, migration, and invasion abilities enhanced
   by CHAC1. More importantly, these malignant phenotypes could be rescued
   by reintroducing PKM2 WT but not PKM2 MT (Fig. [170]6E and
   Supplementary Fig. [171]S6C, D). This indicates that the effects of
   CHAC1 on cell proliferation and metastatic capacity depend on the
   nuclear translocation of PKM2.

   Nuclear PKM2 acts as both a protein kinase and a co-transcription
   factor, promoting the glycolytic process in tumors [[172]20]. We
   initially assessed the correlation between CHAC1 expression and
   downstream glycolytic genes regulated by PKM2. Bioinformatic analysis
   revealed a positive correlation between CHAC1 expression and GLUT1
   (R = 0.174, p < 0.001) and LDHA (R = 0.088, p = 0.040, Supplementary
   Fig. [173]S6E). ENO1 was also correlated with CHAC1 but not
   significantly (Supplementary Fig. [174]S6E). Upon examining the
   expression of glycolytic genes in cells, we observed that CHAC1
   overexpression significantly upregulated the mRNA and protein levels of
   GLUT1, ENO1, and LDHA. Conversely, the knockdown of CHAC1 led to a
   downregulation of these glycolytic genes (Fig. [175]6G, H). Further
   analysis using human LUAD tissue samples demonstrated a significant
   positive correlation between CHAC1 and GLUT1 (R = 0.438, p = 0.018),
   ENO1 (R = 0.869, p < 0.001), and LDHA expression (R = 0.650, p < 0.001,
   Fig. [176]6I). Notably, as LDH catalyzing the conversion of pyruvate to
   lactate, CHAC1 overexpression also resulted in increased LDH activity
   (Fig. [177]6J).

   Given that PKM2 acts as a co-transcription factor, potentially
   interacting with transcription factors (TFs) like c-Myc in the nucleus
   to influence target gene regulation [[178]20], we investigated whether
   c-Myc expression was affected. Our results indicate that CHAC1
   overexpression promotes c-Myc mRNA expression; however, it does not
   significantly alter c-Myc protein levels or its distribution between
   the nucleus and cytoplasm (Supplementary Fig. [179]S6F, G). This
   suggests that nuclear PKM2 may regulate glycolytic gene expression
   through other mechanisms.

   STAT3 phosphorylation has been reported as a known substrate of PKM2
   when functioning as a protein kinase [[180]34, [181]35]. Activated
   STAT3, as a TF, regulates the expression of glycolytic genes such as
   GLUT1, ENO1, and LDHA. Thus, we examined the effect of CHAC1 on the
   phosphorylation of STAT3. Our findings indicated that overexpression of
   CHAC1 increased the phosphorylation of STAT3^Y705, whereas knockdown of
   CHAC1 significantly reduced the levels of phosphorylated STAT3^Y705
   (Fig. [182]6K). The selective STAT3 inhibitor 7a, which blocks
   STAT3^Y705 phosphorylation [[183]36], was co-administered with
   overexpressed CHAC1, resulting in a marked suppression of GLUT1, ENO1,
   and LDHA protein expression (Fig. [184]6L). This confirms that CHAC1’s
   regulation of glycolytic genes is mediated via STAT3^Y705
   phosphorylation.

   We explored the effect of the PKM2 MT on STAT3^Y705 phosphorylation.
   The results showed that PKM2 downregulation blocked the CHAC1-induced
   increase in the phosphorylation of STAT3^Y705 (Fig. [185]6M). Similar
   results were observed for the transcription of PKM2 target genes
   (Supplementary Fig. [186]6H). Collectively, these results suggest that
   nuclear PKM2-mediated glycolysis is critical for CHAC1 to promote LUAD
   progression and metastasis.

E2F Transcription Factor 1 (E2F1) activates CHAC1 transcription in LUAD cells

   To elucidate the mechanism of CHAC1 overexpression in LUAD, we
   performed promoter analysis to investigate whether CHAC1 expression is
   regulated by certain TFs. The PROMO [[187]37, [188]38] and hTFtarget
   [[189]39] databases were used to mine potential regulators, and ten TFs
   with high binding potential were obtained (Fig. [190]7A). Then, JASPAR
   database was utilized to analyze the potential binding of TFs to CHAC1
   [[191]40]. For the JASPAR score, higher scores were observed for YY1,
   MAZ, E2F1, STAT1, YY1, and JUND (score > 8) (Supplementary Table
   [192]S8). Correlation analysis showed that MAZ (r = 0.250, p < 0.001),
   E2F1 (r = 0.305, p < 0.001), and CHAC1 were moderately positively
   correlated (Fig. [193]7B, C). After MAZ and E2F1 knockdown, CHAC1 mRNA
   expression was downregulated (Fig. [194]7D, E).

Fig. 7. E2F Transcription Factor 1 (E2F1) activates CHAC1 transcription in
LUAD cells.

   [195]Fig. 7
   [196]Open in a new tab

   A Screening transcription factors (TFs) that may act on CHAC1 promoter
   by PROMO and hTFtarget. B, C Analyze the correlation between TFs and
   CHAC1 expression. D, E E2F1 siRNA or MAZ siRNA were transfected to
   detect CHAC1 mRNA expression. F Dual-luciferase reporter assay was used
   to analyze the effect of E2F1 or MAZ plasmid transfection on the
   fluorescence activity of CHAC1 promoter. G Detect the protein
   expression of CHAC1 in E2F1 overexpressing or knockdown cells. H
   ChIP-qPCR assay detects the fragments that E2F1 may bind to CHAC1.
   Results are shown as means ± SD. Values below western blot bands are
   optical density of signal. *p < 0.05, ***p < 0.001, ns no statistical
   significance.

   To explore whether there is a direct binding interaction between E2F1,
   MAZ, and the CHAC1 promoter, we inserted a CHAC1-containing promoter
   fragment into a reporter plasmid. The plasmid containing the CHAC1
   promoter was co-transfected with E2F1- and MAZ-overexpression plasmids,
   and E2F1 significantly increased the relative luciferase activity in
   LUAD cells (Fig. [197]7F). Moreover, the overexpression of E2F1
   promoted CHAC1 protein expression, whereas knocking down E2F1 had the
   opposite effect (Fig. [198]7G). JASPAR software predicted two possible
   binding sites for E2F1 on CHAC1 (Fig. [199]7H), and the ChIP-qPCR assay
   showed that the P1 fragment of E2F1 had a strong possibility of binding
   to the CHAC1 promoter (Fig. [200]7H). These results suggested that E2F1
   activates CHAC1 transcription in LUAD.

CHAC1 promotes progression of LUAD in vivo

   To explore the association between CHAC1 and LUAD development in vivo,
   A549 and PC9 cells were transfected with CHAC1 overexpression and
   shCHAC1 lentiviruses, respectively. Subsequently, these cells were
   injected into nude mice to construct a xenograft neoplasm system. As
   revealed by the findings, CHAC1 overexpression significantly increased
   the volume and weight of tumors in the mice (Fig. [201]8A, B) and vice
   versa (Fig. [202]8C). In addition, after the CHAC1 was overexpressed,
   the expression level of the proliferation marker Ki67 was upregulated,
   whereas the knockdown of CHAC1 had the opposite effect (Fig. [203]8D,
   E).

Fig. 8. CHAC1 promotes LUAD growth in vivo.

   [204]Fig. 8
   [205]Open in a new tab

   A–C In vivo growth of indicated cell lines stably overexpression or
   silencing CHAC1 were examined (n = 6). Xenograft volume (cm) and weight
   (g) were measured. Tumor volume was calculated by 1/2
   (length × width^2). Data are expressed as means ± SD. Significant
   difference was based on unpaired t-test. D, E Immunohistochemistry
   staining and IHC scores of CHAC1 and Ki67 in respective xenograft tumor
   tissues. F–H In vivo growth of A549 cells stably co-transfected with
   CHAC1 OE and shPKM2 lentivirus (n = 6). Xenograft volume (cm) and
   weight (g) were measured. I Immunohistochemistry staining of CHAC1,
   PKM2, and Ki67 in respective xenograft tumor tissues. J Working model
   illustrating that CHAC1 promotes glycolysis and progression of LUAD by
   regulating PKM2 SUMOylation. *p < 0.05, **p < 0.01, ***p < 0.001, ns no
   statistical significance.

   We further analyzed whether CHAC1/PKM2 axis in vivo affected orthotopic
   xenograft tumor growth. First, a PKM2-knockdown lentivirus was
   constructed in A549 cells. Among them, the knockdown efficiency of
   shPKM2#1 was the highest, and this lentivirus was used for subsequent
   experiments (Supplementary Fig. [206]S7A). In the subcutaneous
   xenograft model, CHAC1- and shPKM2-expressing cells inhibited LUAD
   orthotopic xenograft tumor growth when compared with CHAC1-expressing
   cells (Fig. [207]8F–H). Moreover, IHC showed that the expression of
   Ki67 in the CHAC1- and shPKM2-expressing groups was lower than that in
   the CHAC1-expressing group (Fig. [208]8I and Supplementary Fig.
   [209]S7B). Taken together, our findings suggest that CHAC1/PKM2 axis is
   essential for LUAD progression.

Discussion

   The biological role and molecular mechanism of action of CHAC1 in LUAD
   remain unclear. In the current study, we demonstrated that the TF E2F1
   activates CHAC1 transcription, resulting in the upregulation of CHAC1
   mRNA and protein expression. CHAC1 functions as a bridge connecting
   UBA2 and PKM2, leading to the induction of PKM2 SUMOylation.
   Furthermore, SUMO-modified PKM2 nuclear translocation activates the
   transcription of proliferation- and glycolysis-related genes, and CHAC1
   promotes LUAD glycolysis and tumor progression.

   CHAC1 plays an important role in determining cell fate. First, CHAC1
   prevents the glutathione peroxidase 4–reduced GSH system, leading to
   the lethal accumulation of lipid peroxides on the cell membrane and
   ferroptosis [[210]41]. This has been reported in tumor and non-tumor
   diseases. CHAC1 acts as a regulatory gene in the “unfolded protein
   response” pathway and regulates apoptosis [[211]42]. Although these
   studies suggest that CHAC1 expression correlates with cell fate, the
   effect of CHAC1 expression on cell fate depends on the cell type. High
   CHAC1 expression is associated with poor prognosis in breast cancer
   [[212]5, [213]6], ovarian cancer [[214]43], renal hyaline carcinoma
   [[215]7], and uveal melanoma [[216]8]. The opposite is true for the
   liver [[217]11], gastric [[218]10], prostate [[219]9], and oral
   squamous cell carcinomas [[220]44]. However, current studies have
   mainly focused on bioinformatic analysis and not in-depth mechanistic
   studies, and the role of CHAC1 in lung cancer progression remains
   unclear. In this study, we demonstrated for the first time that CHAC1
   expression is upregulated in LUAD and is correlated with its malignant
   progression.

   To investigate the mechanism of action of CHAC1 in LUAD, we defined the
   upstream promoter sequence of CHAC1 and confirmed that E2F1 is a TF
   that upregulates CHAC1 in LUAD. E2F1 is a member of the E2F TF family,
   which can activate or silence the transcription of various genes and is
   an important regulatory factor in cell proliferation, differentiation,
   apoptosis, and drug resistance. As the first member of the E2F family,
   E2F1 acts as a transcriptional cofactor and interacts with
   retinoblastoma proteins to regulate the transcription of
   cell-cycle-related proteins [[221]45]. E2F1 also serves as a TF that
   plays an important role in cancer progression. In HCC, E2F1 upregulates
   the RNA-binding protein AUF1, promoting HCC development and
   chemoresistance [[222]46]. In ccRCC, overexpression of E2F1 upregulates
   SREBP1-induced abnormal lipid metabolism, increasing ccRCC
   proliferation and metastasis [[223]47]. Here, we verified that E2F1
   directly binds to the CHAC1 promoter region and activates its
   transcription, thereby upregulating CHAC1 mRNA and protein expression.

   Currently, the underlying molecular mechanism of action of CHAC1 in
   LUAD progression is unclear. We used bioinformatic analysis, RNA-seq,
   and LC–MS to detect signaling pathways and interacting proteins that
   the abnormal expression of CHAC1 may influence. In the present study,
   we report that CHAC1 affects PK activity but not PKM2 protein
   expression. PK is the rate-limiting enzyme in glycolysis and catalyzes
   the conversion of PEP to PE. PKM encodes two PK isoforms, PKM1 and
   PKM2, via alternative splicing [[224]48]. Among these, PKM1 exists only
   as a highly active tetramer, whereas PKM2 switches between that and a
   low-activity dimer [[225]49]. The low catalytic activity of the dimer
   PKM2 increases the production of glycolytic intermediates [[226]49,
   [227]50]. Changes in the PK enzymatic activity are associated with Y105
   phosphorylation [[228]34], alternative splicing of transcripts
   [[229]51], posttranslational modifications [[230]52, [231]53], and
   hypoxia [[232]54]. We demonstrated that CHAC1 can induce the
   SUMOylation of PKM2 and promote the nuclear translocation of PKM2
   dimers. Re-localization of PKM2 to the nucleus suggests it plays a
   crucial role in this process. In our study, nuclear PKM2 acted as a
   protein kinase, promoting the phosphorylation of STAT3, thereby
   regulating the expression of downstream glycolysis-related genes such
   as GLUT1, ENO1, and LDHA, which in turn promotes tumor progression.
   This is consistent with the findings of Zhao et al. [[233]33, [234]34].

   The SAE1/UBA2 heterodimer is involved in SUMOylation and is a key
   kinase that controls SUMO protein activity. SUMOylation is involved in
   glycolytic metabolism, and UBA2 knockdown inhibits SUMO protein
   activation and maintains high PK activity, thereby reducing glycolytic
   metabolism in rheumatoid arthritis [[235]27]. In HCC, UBA2 activates
   SUMO1, promotes PKM2 nuclear translocation, and activates EMT and the
   STAT3 signaling pathway to promote HCC progression [[236]55]. In this
   study, we revealed that CHAC1 interacted with UBA2 but did not affect
   UBA2 expression, and that CHAC1 maintained low PK activity and dimer
   PKM2 nuclear transfer. The phosphorylation of PKM2 at Y105 induces the
   formation of inactive tetrameric PKM2 dimers and decreases PK activity
   [[237]53]. SAE1/UBA2-mediated SUMOylation blocks PKM2 phosphorylation
   and increases PK activity [[238]27]. Therefore, this was not validated
   in this study.

   In this study, the overexpression of CHAC1 caused downregulation of PK
   activity and unchanged PKM2 protein expression. PKM2 may function
   mainly as a protein kinase to regulate gene transcription in the
   nucleus. PKM2 SIM IKII265-268 site mutation eliminates SUMO1-induced
   PKM2 SUMOylation, which affects PKM nuclear translocation [[239]31,
   [240]33]. In the present study, transfection with the PKM2 WT plasmid
   increased PKM2 nuclear translocation, whereas transfection with the
   PKM2 MT plasmid failed to induce PKM2 nuclear translocation.
   Importantly, the PKM2 WT plasmid enhanced the PKM2-dependent Warburg
   effect. Our study provides evidence for a model in which CHAC1 promotes
   PKM2 nonmetabolic enzyme activity by promoting PKM2 SUMOylation and
   nuclear translocation.

   This study conducted an in-depth mechanistic investigation of the role
   of CHAC1 in LUAD; however, there are still limitations to this study.
   Transgenic models with CHAC1 knockouts may better determine the role of
   CHAC1 in LUAD. This model is expected to be used for further
   verification in the future.

   Taken together, we confirm that CHAC1 is a key target of the Warburg
   effect, promoting the malignant progression of LUAD. E2F1 directly
   targets the CHAC1 promoter, efficiently activating CHAC1 transcription
   in LUAD. CHAC1 acts as a bridge connecting UBA2 and PKM2, facilitating
   PKM2 SUMOylation and its translocation from the cytoplasm to the
   nucleus. Additionally, PKM2 functions as a protein kinase, activating
   the STAT3 signaling pathway and upregulating the transcription of genes
   associated with glycolysis, ultimately leading to cell proliferation
   and metastasis (Fig. [241]8J). Our study indicates that targeting CHAC1
   or blocking the CHAC1/UBA2/PKM2 axis may be a promising treatment
   strategy for LUAD.

Supplementary information

   [242]Supplementary Data^ (30MB, docx)
   [243]Supplementary Data^ (1.4MB, pdf)
   [244]Supplementary Data^ (44.3KB, xlsx)
   [245]Supplementary Data^ (11KB, xlsx)

Acknowledgements