CLTRN, regulated by NRF1/RAN/DLD protein complex, enhances radiation sensitivity of hepatocellular carcinoma cells through ferroptosis pathway

Radiotherapy is a viable treatment option for patients with unresectable hepatocellular carcinoma (HCC). However, radiation resistance is an issue that needs to be addressed. In this context, cumulative evidence supports the functional roles of a variety of RNA or proteins in radioresistance, and suggests that the modulation of their expression may constitute a novel radiosensitization approach. Here, we investigated the ability of collectrin (CLTRN) to enhance the radiosensitivity in HCC patients for the rst time. Transcriptome sequencing technology (RNA-seq technology) was used to analyze the transcription-level changes in the genes from the HepG2 cells before and after X-ray irradiation. Combining the results with the HCC tissue RNA-seq data, we determined the ultimate target gene through bioinformatics analysis and cellular verication. A series of cellular and molecular biology techniques were applied in vitro and in vivo to conrm whether CLTRN can enhance radiosensitivity in HCC cells. Subsequently, the downstream action mechanism, the upstream transcription factor, and the interaction proteins of CLTRN were determined.


RNA-seq
Total RNA was extracted using the RNeasy Kit (AM1924, Invitrogen), and further treated with DNase to remove genomic DNA contamination. Isolation of mRNA was performed using the NEB. Next PolyA mRNA Magnetic Isolation Module (New England Biolabs, Ipswich, MA, USA) and the mRNA was then used for RNA-Seq library preparation with the NEB Next Ultra Directional RNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, USA). The library was then subjected to Illumina sequencing with paired-end 2x150 as the sequencing mode. Raw reads were ltered to obtain high-quality clean reads by removing sequencing adapters, short reads (length <100 bp) and low-quality reads using Cutadapt (v1.9.1) and Trimmomatic (v0. 35). Then Fast QC is used to ensure high reads quality. The clean reads were mapped to the mouse genome (assembly GRCm38) using the HISAT2 software. Gene expression levels were estimated using FPKM (fragments per kilobase of exon per million fragments mapped) by String Tie. Ballgown, a R package, were used to measure differential gene expression. The false discovery rate (FDR) control method was used to calculate the adjusted P-values in multiple testing in order to evaluate the signi cance of the differences. Here, only gene with an adjusted P-value <0.05 were used for subsequent analysis.

Immunohistochemistry
Tumors were blocked with peroxide and non-immune animal serum and incubated sequentially with indicated rst antibodies and biotin-labeled goat anti-rabbit IgG (dilution, 1:1000). Eventually, the tumors were stained with Dolichos bi orus aggultinin (DBA), counterstained with hematoxylin, dehydrated, cleared in xylene, and then xed. The antibody information is shown in Table S1.
Quantitative reverse transcription polymerase chain reaction (RT-qPCR) Total RNA was extracted using the RNeasy Kit (AM1924, Invitrogen) according to the manufacturer's instructions, and reversely transcribed into cDNA using AMV reverse transcriptase, which was then used for ampli cation of the purpose genes. Furthermore, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was employed as an internal standard. Primer sequences used for RT-qPCR analysis are listed in Table S2.
Cell culture In the current research, 4 HCC cell lines (Huh-7, MHCC-97H, SMMC-7721, and HepG2) and a normal human liver cell line (LO2) were purchased from the ATCC cell bank. All the cell lines were seeded in 6-well plates at a density of 1.5 ×10 5 cells/well and maintained in Roswell Park Memorial Institute (RPMI)-1640 (A1049101,Gibco) supplemented with 10% fetal bovine serum (FBS). All the cells were cultured at 37 °C in a humidi ed atmosphere containing 5% CO2.

Western blot analysis
In the present study, total protein was extracted from cells and tissues subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Protein concentrations were transferred onto polyvinylidene di uoride (PVDF) membranes, and then membranes were blocked and incubated with primary antibody (dilution, 1:1000) at 4 °C overnight. After washing three times with TBS-T solution for 10 min, the membranes underwent hybridization with a goat anti-rabbit IgG secondary antibody (dilution, 1:1000) at 37 °C for 1 h. After further washing, target protein levels were visualized using an enhanced chemiluminescence kit (No. 35055, Thermo Scienti c). The antibody information is shown in Table S1.

Transfection
For transfection, the cells were starved 2 h prior to transfection, using Lipofectamine 3000 (L3000015,Invitrogen) on the basis of the manufacturers' protocols. Moreover, at 40 h after transfection, cells were treated with complete media for the indicated period before harvesting for further analysis. The transfection product information are shown in Table S3.
Ionizing radiation (IR) and colony formation HepG2 cells were seeded into 6-well plates, then exposed to 4 Gy(1 Gy/ min) of X-ray (Rad source, RS2000) and cultured for 48h. Total RNA in the control group and the radiation group were extracted and sequenced. HepG2and Smmc-7721 cells transfected with NC or clone_CLTRN were seeded into 6-well plates, 4Gy for function experiment and ferroptosis detection. Brie y to colony formation, cells were exposed to a range of radiation doses (0, 2, 4, 6 and 8 Gy) and 12 days after irradiation surviving colonies were stained with 0.5% crystal violet and counted. The survival fraction was calculated as the numbers of colonies divided by the numbers of cells seeded times plating e ciency. The plating e ciency was calculated as the ratio of the number of colonies (consisting of at least 50 cells) to the number of seeded cells. The surviving fraction was calculated as the ratio of the plating e ciency of the irradiated cells to that of the non-irradiated ones.
Cell migration and invasion assays.
Cell migration and invasion assays were performed using Millicell Cell Culture Inserts (24-well plates; 8 µm pore size). Stably transduced cells were used for these assays. For the migration assay, cells in serum-free medium were seeded on the upper chambers. For the invasion assays, the membranes of the upper chambers were coated with 8 µl Matrigel (No.A1413301, Gibco) in 32 µl RPMI-1640 medium for 3 h in a humidi ed incubator. The cells were then seeded in the coated upper chambers. RPMI-1640 medium containing 10% FBS was added into the lower chambers. Cells were incubated for 24 or 48 h for the migration assays, and 48 or 72 h for the invasion assays, respectively. Then, the cells on the lower membranes were stained using a Wright-Giemsa Stain kit (No.9990710, Thermo Scienti c) and observed at x100 magni cation. Five elds were randomly selected and cells were counted upon observation under a light microscope, the number of cells of average per eld was calculated nally.
Apoptosis and cell cycle assay Cells were trypsinized and centrifuged at 1500 rpm/min for 5 min. Cells were harvested and washed with Phosphate Buffered Saline (PBS) twice. Reagents for apoptosis detection were added, and then cells were incubated in dark for 30 min and subjected to ow cytometry analysis (FACS) (Beckman Coulter, Brea, CA, USA). Additionally, cells were collected, washed with PBS, xed with 75% ethanol at -20°C overnight, and centrifuged at 1500 rpm/min for 5 min. Then, ethanol was removed and cells were washed with PBS twice. Propidium iodide (PI) and 500 μl of RNAse were added, and then cells were incubated in dark at 4°C for 60 min. Lastly, cells were subjected to cell cycle analysis by FACS.

Animal Studies
A total of 24 BALB/c male nude mice (8 weeks old) were purchased from Shanghai X-B Animal. Ltd(Shanghai, China). All mice were kept in pathogen-free cages in 26-28˚C and 50% humidity.
Clone_CLTRN cells and control RNA cells were resuspended in 100 µl PBS and subcutaneously injected into the right side of nude mice (3x10 6 cells/mouse, 6 nude mice per group). Next, a 10-Gy dose of IR was delivered to the tumor cell injection site in 9,15days after injection or not. Tumor volumes were measured after 9 days, and every 3 days afterwards. Tumor volume was calculated using the formula: V (mm 3 ) = width 2 (mm 2 ) x length (mm)/2. All the mice were sacri ced 3 weeks after the injection and tumors were harvested for analysis.

Chromatin immunoprecipitation(CHIP)
Cells were xed with formaldehyde 1% for 15 min, then cells and nuclei were lysed. The recovered chromatin was sonicated: 10 cycles of 10s (1s on, 1 s off) and precleared. One hundred micrograms chromatin were used for immunoprecipitation with 2 µg of antibodies of NRF1(ab34682, Abcam). Then, the chromatin was incubated with A/G beads for 2h. Crosslinking was reversed by incubation of the beads with SDS at 70°C and proteins were degraded with proteinase K. Finally, DNA was puri ed using the DNA puri cation kit (No.142095, CST), and CHIP was analyzed by qPCR using speci c primers (Sangon biotech, Shanghai) . Primer sequences are shown in Table S4.

Co-immunoprecipitation (CO-IP)
For the co-immunoprecipitation assay (Co-IP), the cells were lysed with modi ed TNE buffer (50 mM Tris [pH 8.0], 1% Nonidet P-40 [NP-40], 150 mM NaCl, 2 mM EDTA, 10 mM sodium uoride, 10 mM sodium pyrophosphate) supplemented with 1 mg/L aprotinin, 1 mM sodium orthovanadate (Na 3 VO 4 ) and 1 mg/L leupeptin. Lysates were centrifuged and cleared by incubation with 25 μl of Protein A/G gel for 1.5 hr at 4°C. The pre-cleared supernatant was subjected to IP using the indicated rst antibodies at 4°C overnight. Then, the protein complexes were collected by incubation with 30 μl of Protein A/G gel for 2 hr at 4°C. The protein complexes were resolved by SDS-PAGE. Subsequently, western blot was performed. The antibody information is shown in Table S1.
Coomassie Bright Blue staining and mass spectrum identi cation The immunoprecipitation and SDS-PAGE electrophoresis procedures were the same as those for coimmunoprecipitation. The adhesive was placed in a glass dish, Coomassie Bright Blue solution was added, the solution was incubated at 70 °C for 5 min and placed on a shaker for 30 min. The Coomassie Bright Blue solution was then recycled, the adhesive was rinsed with clean water, the Coomassie Bright Blue decolorizing agent was added, and the solution was incubated overnight in a decolorizing shaker. The decolorizing liquid was removed and the adhesive was rinsed with clean water and recovered for using with the target strip. A protein gel recovery kit (No.C500062-0020, Sangon Biotech) was used to recover the target protein. Detection and identi cation were conducted using mass spectrometry.

Reactive oxygen species (ROS) assay
After trypsin digestion, the cells were washed with PBS and the number of cells was adjusted to 1×10 -6-2×10 8 cells. DCFH-DA was diluted with serum-free medium at a 1:1000 ratio to a nal concentration of 10 µmol/L. To each tube of the sample, 1 mL of the staining solution was added, and the suspension was incubated at 37 °C for 20 min with upside-down mixing every 3-5 min to establish complete contact between the probe and the cells. The cells were washed with PBS and observed using ow cytometry.

Transmission electron microscopy
After the culture medium was discarded, the electron microscope xator was added and the cells were incubated for 2-4 h at 4 ºC. The cells were centrifuged at a low speed to precipitate them to the bottom of the tube, and cells in the size of mung beans were observed. The cell pellets were wrapped in 1% agarose and rinsed thrice with 0.1 M phosphate buffer (PB, pH 7.4), for 15 min in each round. Next, 1% osmium in 0.1 M PB (pH 7.4) were used to x the cells at room temperature (20 ºC) and the culture was maintained for 2 h, followed by rinsing with PB (pH 7.4) thrice for 15 min in each round. The tissues were successively dehydrated with 50-70-80-90-95-100-100% alcohol-100% acetone -100% acetone, for 15 min at each step. The following sequential treatment protocol was adopted for the cells: acetone: 812 embedding agent at 1:1 ratio for 2-4 h; acetone: 812 embedding agent at 1:2 ratio, allowed to permeate overnight; treatment with pure 812 embedding agent for 5-8 h. The pure 812 embedding agent was then transferred to the embedding plate, and the sample was placed on the embedding plate and incubated overnight at 37 ºC. Embedment, sectioning, uranium lead double staining followed after this step, after which the samples were observed under a transmission electron microscope.

Iron Assay
The relative iron concentration in cell lysates was assessed using the Iron Assay Kit (No. E-BC-K139-M; Elabscience) according to the manufacturer's instructions.

Immuno uorescence
Cells were grown on poly-L-lysine-coated glass coverslips (BD Biosciences, San Jose, CA), and then xed with 4% paraformaldehyde, and permeabilized with PBS containing 0.1% Triton X-100 (PBS-T). Coverslips were incubated in blocking solution containing 2% BSA in PBS for 1 h, and incubated with the appropriate primary RAN and DLD antibody for 1 h at room temperature. After incubation with Alexa Fluor 594conjugated (red) goat anti-rabbit secondary antibody, cells were stained with DAPI for nuclear staining and then visualized by uorescence microscopy. The antibody information is shown in Table S1.

Statistical analysis
In the present research, SPSS 19.0 software (IBM, Armonk, NU, USA), GraphPad Prism 6.0 software (GraphPad Software, La Jolla, CA) and R3.6.1 (AT&T Bell Laboratories, USA) were used to perform statistical analyses. Quantitative data were expressed as mean ± standard deviation (SD), and were examined by independent sample t-test. The counting data were analyzed by chi-square test. The Kaplan-Meier estimator was used to estimate the patients' overall survival rates. P<0.05 was considered statistically signi cant.

Results of RNA-seq and identi cation of target genes
In this study, we studied three pairs of genes from HCC cells and paracancer tissues (BioProject:PRJNA642330)(FigS1A-B). A search in the GEO database yielded 23 pairs of raw data, and not transcripts, under the project PRJNA576155. The cumulative distribution curve from the analysis indicated that the difference was highly consistent [9](Fig1A, FigS1C, E). In addition, we analyzed the changes in protein-coding genes at the transcription level in HepG2 cells before and after X-ray irradiation(BioProject:PRJNA642330) (Fig1B, FigS1D). We screened data based on the following conditions: 1. Upregulated expression in HCC specimens and downregulated expression in cell lines after X-ray irradiation; 2. Downregulated genes in HCC specimens and upregulated genes in cell lines; 3. Q value ranks of the top 30 genes in both groups. Thirteen genes were obtained, and we studied these to further identify the gene of interest in this study(Fig1C-D). We used the QPCR method to analyze transcription level changes in the 13 genes from HepG2 cells before and after X-ray irradiation, and eventually selected CLTRN as the gene of interest, based on the fact that it had undergone the most signi cant changes(Fig2A).

CLTRN is expressed at low levels in HCC cell lines and tissues, and inhibition of CLTRN expression is associated with poor survival in HCC
The database revealed that CLTRN is primarily localized to the membrane in HCC cells, and we veri ed the same in HepG2 cells (Fig2B). Thirty cases of paired HCC tumor and paracancer tissue samples were selected to study the expression levels of CLTRN in each tissue sample using RT-qPCR. The results indicated that the expression level of CLTRN in tumor tissues was signi cantly lower than that in normal tissues (Fig2D). In addition, huh-7, mhcc-97h, smmc-7721, HepG2, and the normal stem cell line LO2 were selected for the analysis of CLTRN expression levels. We observed that the expression level of CLTRN in tumor cells was lower than that in normal cells, particularly in HepG2 cells (Fig2C).The expression levels of CLTRN in 109 tumor tissues from selected biobanks were analyzed using RT-PCR and accordingly, the samples were categorized into high expression and low expression groups according to the median value. Kaplan-Meier analysis revealed that low expression levels of CLTRN corresponded to poorer survival and higher probability of recurrence (Fig2E-F). In the current study, the Cox proportional-hazards model was used to analyze the clinical data and prognosis in HCC patients. We determined the expression level of CLTRN, TNM staging, and vascular invasion to be independent clinical predictors of survival in HCC, whereas tumor encapsulation, LRRN3 expression, and vascular invasion were independent clinical predictors of recurrence time in HCC (Table 2).

CLTRN overexpression can increase the radiosensitivity of HCC cells in vitro
To investigate the effect of CLTRN on radiosensitivity in HCC cells, HepG2 and SMMC-7721 cells were CLTRN overexpression can reduce tumorigenicity in nude mice post-irradiation HepG2 cells with stable overexpression of CLTRN were established for studying in vivo tumorigenesis in nude mice. We randomly divided 24 nude mice into four groups, namely NC, clone_CLTRN, NC+IR, and clone_CLTRN+IR. The tumor was treated using radiation during tumor growth, and the tumor formation potential in each group was compared. We observed that the tumor volume was the lowest in nude mice that were treated with radiation after CLTRN overexpression, which indicates that CLTRN overexpression could enhance the radiosensitivity of HCC cells (Fig4A-C).
CLTRN is primarily associated with glutathione metabolism in HCC cells.
We used RNA-seq to analyze the changes in protein-coding genes at the transcription level induced upon CLTRN overexpression in HepG2 cells (BioProject:PRJNA642330) (Fig5A-B). Combining our data with those from KEGG database and bioinformatics analyses, we observed that the altered genes were mostly enriched in the glutathione metabolic pathway (Fig5C). We hypothesized that CLTRN might be associated with glutathione metabolism.
CLTRN overexpression can increase ferroptosis in HCC, and CLTRN can enhance the radiosensitivity of HCC cells through the ferroptosis pathway.
Since glutathione metabolism is a key mechanism in ferroptosis, we further hypothesized that CLTRN could play a role in the ferroptosis pathway (Fig5D) [10]. We performed the relevant validation experiments in HepG2 and SMCC-7721 cells. The cells were divided into the following three groups: NC, NC+IR, and clone_CLTRN+IR. The concentration of iron ions in the cell lysate and the levels of ROS in the liposomes were studied. Lastly, the changes in the organelles in each group were observed using transmission electron microscopy. We found that radiation could independently induce an increase in the concentration of cellular iron ions and could also increase the levels of ROS in cellular liposomes (Fig6A-B), which suggests that radiation can induce ferroptosis in HCC. The results of transmission electron microscopy also revealed that the number of mitochondria decreased, their membrane density increased, and the number of cristae in mitochondria decreased in cells from the NC+IR group (Fig6C); however, compared to the cells from the NC+IR group, the radiosensitivity of tumor cells increased signi cantly upon the overexpression of CLTRN (Fig6C). The above experimental results indicated that CLTRN enhanced the radiosensitivity of tumor cells by increasing ferroptosis. Furthermore, we compared the changes in the expression of protein indicators associated with glutathione metabolism and ferroptosis, including glutathione peroxidase 4 (GPX4), solute carrier family 7 member 11 (SLC7A11), ferritin, heavy polypeptide 1 (FTH1), prostaglandin endoperoxide synthase 2 (PTGS2), NADPH oxidase 1 (NOX1), and acyl-CoA synthetase long-chain family member 4 (ACSL4), after CLTRN overexpression. The results indicated that PTGS2, NOX1, and ACSL4 were upregulated, while GPX4, SLC7A11, and FTH1 were downregulated in cells that overexpressed CLTRN, which further con rmed the function of CLTRN in the ferroptosis pathway and its effect on the proliferation of tumor cells (Fig6D). Moreover, the levels of GPX4 and SLC7A11 expression were also the lowest in the clone_CLTRN+IR group, as observed from the immunohistochemical sections of tissue samples from nude mice, which indicates that ferroptosis was most pronounced in this group (Fig4D).
NRF1 is a transcription factor and upstream regulator of CLTRN.
We further investigated the upstream regulatory mechanism of CLTRN. Bedtools was used to extract the sequences 1500 bp upstream and 500 bp downstream of the transcription start site to predict the transcription factor binding site on the JASPAR website. Eleven candidate transcription factors were obtained at p < 0.01 (Fig7A). We determined the RNA expression levels of these 11 candidate transcription factors in HCC and paracancer specimens, and observed that YY1 transcription factor (YY1), HNF1 homeobox A (HNF1A), paired box 5 (PAX5), and NRF1 were downregulated in the tumor and upregulated in paracancer tissues, which was also consistent with the expression pattern of CLTRN (Fig7B). Subsequently, interfering sequences were designed for these four genes and the interference was implemented successfully in HepG2 cells. Western blot analysis of the interfering proteins and CLTRN was conducted in each group. The results indicated that CLTRN expression reduced signi cantly only upon interference with NRF1, which indicates that NRF1 might be the upstream transcription factor of CLTRN (Fig7C-E). We then collected the chip-SEQ data of NRF1 in HepG2 cells (GEO:GSE96424) and analyzed the binding DNA, and found that CLTRN was one of its binding DNA sequences (Fig7G). Using chip-QPCR, we nally con rmed the binding between NRF1 and CLTRN at the site predicted by the software (Fig7F,H).
Member RAS oncogene family (RAN) and Dihydrolipoamide dehydrogenase (DLD)are interacting proteins of NRF1 in HCC cells We attempted to identify the interaction proteins of NRF1 in HCC. We used HepG2 cells as the research object and performed immunoprecipitation with an NRF1 antibody. After Coomassie Blue staining, the samples were electrophoresed, the corresponding fragments were cut out from the gel, and mass spectrometry analysis was conducted (Fig8A-B). For the top 100 proteins obtained using mass spectrometry, we conducted correlation screening in the GEPIA database and selected proteins with R values > 0.3. We attempted to determine the location of these proteins from the UniProt database, selected the nucleus-localized proteins, and identi ed RAN and DLD as the target proteins (Fig8C-D). The immuno uorescence assay revealed that RAN and DLD were indeed located in the nuclei of HepG2 cells (Fig8E). Finally, in the co-immunoprecipitation experiment, we con rmed that RAN and DLD are the interacting proteins of NRF1 (Fig8F).

Discussion
HCC is one of the most common and aggressive malignancies with global prevalence, and ranks sixth in terms of the number of cases and third in terms of cancer mortality [11,12]. Since most patients with HCC do not exhibit speci c symptoms at an early stage, the proportion of patients with resectable HCC is low. Non-surgical treatment is of great importance for patients with HCC who cannot undergo surgical resection. Radiotherapy is an important non-surgical treatment method for cancer; however, its clinical application is limited owing to the prevalence of radiotherapy resistance in patients. Therefore, we primarily focused on methods for improving radiosensitivity in HCC cells.
Here, we rst used RNA-seq technology to study the upregulated genes in HCC tissue chips, along with the genes with altered expression after radiation treatment of HepG2 cells, and found 13 genes that may be associated with HCC radiosensitivity, which was subsequently veri ed. We concluded that CLTRN may function as a HCC radiosensitivity gene.
The collectrin gene (CLTRN, Gene Bank ID:AF178085), also known as Tmem27, was rst isolated and cloned from kidney tissues in 1999 [13]. CLTRN is a type Ia transmembrane glycoprotein located in the cell membrane and is highly conserved among species [14,15]. In mice, collectrin de ciency causes profound aminoaciduria involving both neutral and charged amino acids [16]. The expansion in the biochemical phenotype over that seen in Hartnup disease is thought to be associated with the additional role played by collectrin in the functioning of other renal amino acid transporters [17].
Currently, there are a limited number of studies on the function of CLTRN, and the major ndings are related to its role in the regulation of arterial pressure, nervous system, and blood glucose levels [17][18][19]. CLTRN has also been studied inextensively in tumors (only in kidney cancer) [20]. In this study, we observed that CLTRN can inhibit invasion, migration, and proliferation of HCC cells after irradiation in vitro, which suggests that CLTRN is associated with the biological behavior of liver cancer cells. However, CLTRN did not exert signi cant effect on the cell cycle and apoptosis of HCC cells. In vivo tumor formation in nude mice also revealed that CLTRN could alter the effect of radiation on tumor formation in nude mice, and when overexpressed, it could improve the sensitivity of tumors to radiation. CLTRN is associated with tumor biological behavior in HCC; however, the mechanism underlying the regulation of HCC cells by CLTRN remains unclear. We observed that CLTRN was overexpressed in HepG2 cells, and transcriptional-level changes of the protein-coding genome of the overexpressed and control cells were analyzed using RNA-sEq. KEGG pathway analysis of the transcriptome data revealed that CLTRN was signi cantly associated with glutathione metabolism. Glutathione is widely found in animals and plants and plays an important role in organisms. Furthermore, glutathione metabolism has been linked to ferroptosis.
Ferroptosis, an iron-dependent form of programmed cell death that is induced by excessive lipid peroxidation, is morphologically and mechanistically distinct from apoptosis [21][22][23]. GPX4, a type of glutathione peroxidase, utilizes reduced glutathione to convert lipid hydroperoxides to lipid alcohols, and thereby mitigates lipid peroxidation and inhibits ferroptosis [24][25][26]. The amino acid transporter SLC7A11 (also known as xCT) assists in the delivery of extracellular cysteine into the cell [27,28]. Resultantly, the inactivation of GPX4 or SLC7A11 through genetic or pharmacologic means induces ferroptosis. Recent studies have revealed that FSP1 plays an important role in ferroptosis. A protein similar to other Ufm1speci c proteases (FSP1), a CoQ oxidoreductase, plays a role similar to GPX4 during ferroptosis and has been denoted as a new ferroptosis effector protein, similar to GPX4 [29,30].
We hypothesized that CLTRN is associated with ferroptosis and in uences the tumor biology of HCC cells through ferroptosis. We investigated whether CLTRN overexpression causes biological changes in cellassociated ferroptosis. The results indicated that the cytoplasm of HepG2 cells that overexpress CTLRN underwent signi cant changes in ferroptosis morphology after IR. In addition, compared to the blank control cells, the IR cells also exhibited certain signs of ferroptosis, which suggests that radiation could independently cause ferroptosis in tumor cells. Recent studies have revealed that IR induces ROS generation as well as the expression of ACSL4, which is an lipid metabolic enzyme essential for ferroptosis, and consequently increases lipid peroxidation and ferroptosis. ACSL4 de ciency can largely eliminate IR-induced ferroptosis and enhance radiation resistance [31]. At the protein level, after the overexpression of CLTRN, the expression levels of index proteins associated with glutathione metabolism and ferroptosis alteredsigni cantly, which further indicates that CLTRN plays a role in enhancing the radiosensitivity of tumor cells through glutathione metabolic-ferroptosis pathway.
We identi ed NRF1 as the upstream regulatory transcription factor of CLTRN using database prediction and veri cation. NRFl is a transcription factor encoded by nuclear genes, which regulates the expression of mitochondrial respiratory chains [32]. NRF1 induces the expression of the protein components of the mitochondrial respiratory chain, regulates the expression of mitochondrial transport proteins and mitochondrial oxidative stress-related proteins, and participates in estrogen-mediated mitochondrial biosynthesis and cellular metabolism [33]. NRF1 is associated with the incidence and development of breast cancer, and its dynamic transcriptional regulation patterns can be studied to predict cancer risk and e cacy [34]. Our study is the rst to demonstrate that NRF1 affects the ferroptosis network of cells by regulating CLTRN expression in liver cancer. We predicted Ran and DLD to be the interacting proteins using immunoprecipitation-mass spectrometry in combination with database analysis, and veri ed the same in co-immunoprecipitation experiments. RAN is a GTPase associated with the RAS family and encodes the Ras-related nuclear protein, which is a unique member of the Ras superfamily of GTPases [35]. Mutations in the RAN gene likely play a critical role in pathology-related changes in miRNA transport and expression, and therefore, contribute to tumorigenesis and development [36]. DLD is associated with cysteine metabolism and regulation of ROS generation [37], and its role in tumor growth regulation mediated through the ferroptosis pathway in neck tumors has been investigated [38]. We observed that DLD and RAN regulate CLTRN expression in HCC cells by interacting with NRF1, and thereby affect ferroptosis and enhance the radiosensitivity of HCC cells.

Abbreviations
The JASPAR website predicts the binding sites of NRF1 and CLTRN. g Chip-seq data of NRF1 in the HepG2 cell line showed that NRF1 binds to CLTRN DNA. h NRF1 binding to CLTRN at the predicted sites was veri ed in HepG2 cells. The data are expressed in terms of mean ± SD (Student's t-test; ** P < 0.01; *** P < 0.001;****P < 0.0001; NS no signi cance). lines after interference with YY1, HNF1A, PAX5, and NRF1. When NRF1 activity was impaired, CLTRN expression was signi cantly inhibited, which was different from that observed for the other three genes. f The JASPAR website predicts the binding sites of NRF1 and CLTRN. g Chip-seq data of NRF1 in the HepG2 cell line showed that NRF1 binds to CLTRN DNA. h NRF1 binding to CLTRN at the predicted sites was veri ed in HepG2 cells. The data are expressed in terms of mean ± SD (Student's t-test; ** P < 0.01; *** P < 0.001;****P < 0.0001; NS no signi cance).