A total of 59 samples from patients with breast cancer (tumor tissues and paired normal adjacent tissues) were retrospectively obtained from the Department of Breast Surgery, West China Hospital, Sichuan University (Chengdu, China) between March 2019 and June 2020). The primary cancer tissues and the paired normal adjacent tissues were immediately snap-frozen in liquid nitrogen and stored at −80°C until use for mRNA and protein analyses. The following exclusion criteria were used: i) Patients with incomplete pathological and clinical information; ii) patients with recurrent disease; iii) patients that had received preoperative radiation or chemotherapy, or any other treatment; iv) patients with other preoperative complications; v) patients with cognitive or mental disorders; and vi) pregnant women. The following inclusion criteria were used: i) Breast cancer tissue samples were all pathologically and confirmed as breast cancer. All patients provided written informed consent to participate in the study. The present study was approved by the Institutional Review Board of West China Hospital, Sichuan University (Chengdu, China).

Human BC cell lines, including T47D, MDA-MB-231, MCF-7 and BT-474, and the human normal mammary epithelial cell line MCF-10A, were purchased from the American Type Culture Collection. T47D, MDA-MB-231, MCF-7 and BT-474 cells were cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) containing 10% FBS (Biological Industries). MCF-10A cells were cultured in Minimum Essential Medium (cat. no. CC-3151; Lonza Group, Ltd.) supplemented with Mammary Epithelial Cell Growth Medium, MEGM^® SingleQouts^® (cat. no. CC-4136; Lonza Group, Ltd.) and 10% FBS (Biological Industries). Cells were cultured in a humidified atmosphere containing 5% CO₂ at 37°C.

Cells were transfected with the following plasmids: Overexpression TRIM35 (Oe-TRIM35), short hairpin (sh)RNA targeting TRIM35 for knockdown (Sh-TRIM35) and the respective negative controls (NCs) (Shanghai GenePharma, Co., Ltd.). Plasmids were transfected into breast cancer cells with Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.). The shRNA sequence De-TRIM35 was constructed into the pLVX-shRNA2-BSD lentiviral vector (Takara Bio USA, Inc.). The target sequences of the shRNAs were as follows: TRIM35 1, 5′-CTT GCA TCT GTG GAA TCT GTA-3′; TRIM35 2, 5′-AGT GTC AAG GAA GAA CTG GAT-3′; TRIM35 3, 5′-GCG CTG GAC CAG CTA CCG CTT-3′ (this sequence was used to produce the main results); NC shRNA, 5′-TTC TCC GAA CGT GTC ACG TAA-3′. TRIM35-overexpressing lentiviruses (cat. no. CD527A-1; System Biosciences, LLC) were constructed into the lentiviral vector pCDH-EF1α-MCS-T2A-Puro (System Biosciences, LLC). For the NC group, empty vector was used as instead of the overexpression vector.

In another experiment, the cells were cultured in DMEM (10% FBS) containing the pyruvate kinase M2 (PKM2) activator TEPP-46 (5 µM) (cat. no. ML-265; Med Chem Express) for 24 h at 37°C as directed by the manufacturer.

Total RNA was extracted from breast cancer tumor tissues, paired normal adjacent tissues and cell lines using TRIzol reagent (Takara Bio, Inc.). The extracted RNA was reverse transcribed to cDNA with PrimeScript RT Reagent (Takara Bio, Inc.) according to the manufacturer's instructions. Real-time qPCR was performed with a Prime-Script^® RT Reagent Kit (Takara Bio, Inc.) according to the manufacturer's instructions and a LightCycler system (Roche Diagnostics) was used for detection with the following thermocycling conditions: Initial denaturation step at 95°C for 3 min, followed by 45 cycles of denaturation at 95°C for 20 sec, annealing at 60°C for 30 sec and extension at 72°C for 30 sec. GAPDH was used as an internal control. Relative RNA expression levels were calculated by the 2^−ΔΔCq method (24). The primer sequences were as follows: TRIM35 forward, 5′-CAT CGC CAA GCA CAA TCA GG-3′ and reverse, 5′-GCG TTT TCG GCT CTT GTG TT-3′; GAPDH forward, 5′-GGA GCG ACA TCC GTC CAA AAT-3′ and reverse, 5′-GGC TGT TGT CAA TCT TCT CAT GG-3′.

Total protein was isolated from breast cancer tumor tissues, paired normal adjacent tissues and cell lines using protein extraction buffer (RIPA lysis buffer; Beyotime Institute of Biotechnology). The nuclear protein was extracted using the nuclear protein extraction kit (cat. no. P0027; Beyotime Institute of Biotechnology) according to the manufacturer's instructions. The protein concentration was quantified using a BCA kit (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. Total protein was separated by 10% SDS-PAGE (20 µg protein per lane) and transfer onto a PVDF membrane (MilliporeSigma). The membranes were blocked for 1 h at room temperature with 5% nonfat milk powder (Beyotime Institute of Biotechnology). To detect the expression of PKM2 protein in different conformations, the total protein extracted from the sample could was not heat-denatured. Non-denatured gel sample loading buffer (Beyotime Institute of Biotechnology) was used for protein denaturation. In addition, protein electrophoresis was performed using native PAGE running buffer (Beyotime Institute of Biotechnology). The remaining experimental steps were the same as those of conventional western blot experiments. The membranes were probed at 4°C overnight with antibodies against TRIM35 (cat. no. ab272582; 1:5,000 dilution; Abcam), PKM2 (cat. no. 15822-1-AP; ProteinTech Group, Inc.), Bax (cat. no. ab32503; 1:1,000 dilution; Abcam), Bcl-2 (cat. no. ab182858; 1:2,000 dilution; Abcam), N-cadherin (cat. no. ab76011; 1:5,000 dilution; Abcam), Ki-67 (cat. no. ab16667; 1:5,000 dilution; Abcam), E-cadherin (cat. no. ab40772; 1:10,000 dilution; Abcam), twist1 (cat. no. ab175430; 1:1,000 dilution; Abcam) and β-actin (cat. no. ab8226; 1:5,000 dilution; Abcam). The membranes were then incubated with the appropriate secondary antibodies (1:5,000 dilution; cat. nos. bs-40296G-HRP and bs-40295G-HRP; Bioss) for 1 h at room temperature. Finally, protein expression was analysed by chemiluminescence reagents (Hyperfilm ECL).

The UALCAN database (http://ualcan.path.uab.edu/cgi-bin/ualcan-res.pl) was used to analyze the expression of TRIM35 in the TCGA and CPTAC datasets.

The Search Tool for the Retrieval of Interacting Genes and proteins (STRING; https://cn.string-db.org) was used to analyze the relationship between TRIM35 and PKM2.

Cells were transfected and then inoculated into 96-well plates (2×10³ cells/well). After incubation for the indicated durations, CCK-8 assay reagent (Dojindo Molecular Technologies, Inc.) was added to each well containing cells and plates were incubated according to the manufacturer's instructions. The absorbance at 450 nm was recorded for each well using a microplate reader to determine cell proliferation.

Cell invasion and migration assays were performed using Transwell chambers (8 µm pore size; Costar; Corning, Inc.). MDA-MB-231 and MCF-7 cells (1×10⁵ cells/well) were seeded into the upper chamber of the Transwell chambers with high-glucose DMEM containing 1% FBS. Furthermore, 500 µl of high-glucose DMEM containing 10% FBS was added to the matched lower chamber. After incubation for 48 h, various proportions of MDA-MB-231 and MCF-7 cells had moved to the lower chamber. The cells were fixed with methanol at room temperature for 30 min and stained with 0.1% crystal violet at room temperature for 20 min. For the invasion assay, the inserts were precoated with Matrigel^® (1 mg/ml; Corning, Inc.).

For cell apoptosis, after transfection, MDA-MB-231 and MCF-7 cells were seeded into 6-well plates. Once they reached confluence (60-70%), cells were collected and incubated with Annexin V-FITC (5 µl) (Biogot Technology Co., Ltd.) and propidium iodide (PI) solution (5 µl; Biogot Technology Co., Ltd.) at room temperature for 15 min according to the manufacturer's instructions. Cells were subsequently suspended in 400 µl binding buffer (Biogot Technology Co., Ltd.). Cell apoptosis progression was analyzed using flow cytometry (FACSAria; BD Biosciences). All data were analyzed with ModFit version 4.0 (Verity Software House, Inc.)

After transfection for 36 h, MDA-MB-231 and MCF-7 cells were incubated in DMEM without any L-glucose or phenol red for 8 h. The amount of glucose in the media was measured using a Glucose Colorimetric Assay kit (cat. no. K606-100; BioVision, Inc.) according to the manufacturer's instructions. Fresh DMEM (excluding L-glucose and phenol red) was used for the negative control. A total of three biological replicates were performed.

After transfection for 36 h, MDA-MB-231 and MCF-7 cells were incubated in phenol red-free DMEM without FBS for 4 h. Subsequently, a pyruvate assay kit (cat. no. BC2205), lactate assay kit (cat. no. BC2235) and ATP assay kit (cat. no. BC0300; all from Beijing Solarbio Science & Technology Co., Ltd.) were used to measure intracellular concentrations of pyruvate, lactate and ATP, respectively, according to the manufacturer's instructions. The relative absorbance was measured using a spectrophotometer (Thermo Fisher Scientific, Inc.). The content of pyruvate, lactate and ATP were calculated according to the product manual.

Female BALB/c-nu mice (total number, n=20; age, 5 weeks; weight, 20-25 g) were purchased from the Shanghai Experimental Animal Center and kept at the Experimental Animal Center of West China Hospital of Sichuan University (Chengdu, China). All animal experiments were performed in accordance with the institutional guidelines and were approved by the Committee on the Ethics of Animal Experiments of West China Hospital of Sichuan University (Chengdu, China). For the in vivo xenograft assay, MDA-MB-231 and MCF-7 cells (5×10⁶) transfected with TRIM35 overexpression vector or control were subcutaneously injected into the mammary fat pads of the female athymic nude mice (n=5 in each group). Tumor growth was recorded once every four days with caliper measurements. Tumor volume was calculated according to the following formula: Volume=(width² × length)/2. At 30 days after injection, the mice were euthanized and tumors were collected for analysis. For this, the mice were anesthetized with 1% pentobarbital sodium (45 mg/kg, intraperitoneal) and subsequently euthanized by cervical dislocation (when the heart had stopped completely, the mouse was determined to be dead). Body weight loss >20% was considered a humane endpoint for euthanasia.

Cells were collected and lysed using RIPA Lysis Buffer (cat. no. P0013B; Beyotime Institute of Biotechnology) containing protease inhibitor (cat. no. P1011; Beyotime Institute of Biotechnology). Subsequently, 3 µg of antibody was added to the lysate and the mixture was incubated overnight at 4°C. Protein A/G PLUS-Agarose beads (20 µl; cat. no. sc-2003; Santa Cruz Biotechnology, Inc.) were then added to the mixture and it was incubated on a rotator for 4 h at 4°C. The bead-antibody-protein complexes were washed with precooled PBS three times and then boiled for subsequent western blot analysis (as described above). The antibodies used were as follows: PKM2 (cat. no. 15822-1-AP; 1:1,000 dilution; ProteinTech Group, Inc.) and TRIM35 (cat. no. ab272582; 1:1,000 dilution; Abcam).

MDA-MB-231 and MCF-7 cells were lysed in 1% SDS-containing RIPA buffer by sonication on ice. Lysates were then treated with Protein A/G Plus-Agarose (cat. no. sc-2003; Santa Cruz Biotechnology, Inc.) for 1 h at room temperature. Subsequently, each sample was incubated with IgG (cat. no. 30000-0-AP; 1:5,00 dilution; ProteinTech Group, Inc.) overnight at 4°C. The nuclear pellet was collected by centrifugation at 10,000 × g for 5 min at 4°C and subsequently washed four times using Protein A/G Plus-Agarose beads (by absorbing the protein, washing the beads and then releasing the protein again from the beads). The purified proteins were separated by gradient SDS-PAGE (as in the protocol for western blot analysis). Anti-PKM2 (cat. no. ab137852; 1:1,000 dilution; Abcam) or anti-ubiquitin antibody (cat. no. ab7780; 1:5,00 dilution; Abcam) was used for immunoblotting according to the protocol for western blot analysis.

Tissues were fixed in 10% formalin for 12 h at room temperature and embedded in paraffin for 8 h at room temperature. The slice thickness of the tissue specimen is 4um. The tissue specimen slides were deparaffinized twice with xylene for 15 min each at room temperature and rehydrated three times in a descending series of ethanol solutions (100, 100, 95 and 80%). Antigen retrieval was performed in 10 mmol/l sodium citrate solution (pH 6.0) at 100°C for 16 min and the samples were then cooled for 30 min. Endogenous peroxidase activity was quenched using 3% hydrogen peroxide and blocking non-specific binding was performed with 5% goat serum (OriGene Technologies, Inc.) for 30 min at room temperature. Subsequently, the slides were incubated with anti-N-cadherin (cat. no. ab76011; 1:100 dilution; Abcam), anti-twist1 (cat. no. ab175430; 1:1,000 dilution; Abcam), anti-Ki-67 (cat. no. ab16667; 1:200 dilution; Abcam), anti-E-cadherin (cat. no. ab40772; 1:500 dilution; Abcam), anti-BCL-2 (cat. no. ab182858; 1:500 dilution; Abcam) and anti-BAX (cat. no. ab32503; 1:100 dilution; Abcam) at 4°C overnight. Following the primary antibody incubation, slides were incubated with an anti-rabbit secondary IgG antibody (1:100; cat. no. SAP-9100; OriGene Technologies, Inc.) at 37°C for 30 min. The samples were stained with 3,3-diaminobenzidine and Mayer's hematoxylin. Slides were visualized using a light microscope (magnification, ×100 or ×200; Zeiss AG). Immunostaining results were evaluated by two blinded pathologists. Staining was scored according to the percentage of positive cells and staining intensity. The scoring for staining intensity was as follows: 0, negative; 1, weak; 2, moderate; and 3, strong. The scoring system for the percentage of positive cells was as follows: Negative, 0-5%; 1, 6-25%; 2, 26-50%; 3, 51-75%; and 4, 76-100%. Sections with a total combined score of <4 were classified as having low expression, while sections with a score of ≥4 were considered to have high expression. From each tissue sample, 5 regions were randomly selected for evaluation.

Cells were grown on coverslips that were coated with lysin. Cells (60% confluency) were fixed with 4% paraformaldehyde for 30 min at room temperature and permeabilized with 0.25% Triton X-100 solution for 20 min at room temperature. Cells were washed with PBS and then blocked in 5% bovine serum albumin (Beyotime Institute of Biotechnology) for 1 h at room temperature. Coverslips were incubated with antibodies against PKM2 (cat. no. 15822-1-AP; 1:1,000 dilution; ProteinTech Group, Inc.) overnight at 4°C. After washing with PBS, cells were incubated with appropriate secondary antibody (cat. no. ab150077; 1:5,00 dilution; Abcam) for 1 h at room temperature and DAPI for 10 min at room temperature. Slides were imaged using an inverted fluorescence microscope.

All data were analyzed using SPSS 22.0 software (IBM Corporation) or GraphPad Prism version 7.0 (GraphPad Software, Inc.). Values are expressed as the mean ± standard deviation. Statistical differences (two groups) were analyzed by an unpaired Student's t-test. A paired Student's t-test was used to analyze the statistical difference of TRIM35 expression in breast cancer tissues and adjacent tissues. ANOVA was used for multiple-group comparisons and Tukey's test was used as the post-hoc test after ANOVA. P<0.05 was considered to indicate a statistically significant difference.

Analysis of the TCGA and GEPIA datasets indicated that TRIM35 was low in breast cancer (Fig. 1A and B). In tissue samples collected from patients with breast cancer (age range, 27-76 years; all patients were female; demographic and clinicopathological data provided in Table SI), western blot and RT-qPCR analyses suggested that the expression of TRIM35 in breast cancer tissues was lower than that in adjacent tissues (Fig. 1C and D). Furthermore, the expression level of TRIM35 in breast cancer cell lines was significantly lower than that in the normal breast cell line (Fig. 1E). Through the analysis of the relationship between the expression of TRIM35 and the prognosis of patients with breast cancer in the GEO and TCGA databases, it was revealed that those patients with low TRIM35 expression exhibited significantly shorter overall survival than those with high TRIM35 expression (Fig. 1F and G).

To further verify the role of TRIM35 in breast cancer malignancy, gain-of-function experiments were performed in MDA-MB-231 and MCF-7 cell lines. It was confirmed that TRIM35 was effectively overexpressed in MDA-MB-231 and MCF-7 cell lines (Fig. 2A and B). The CCK-8 assay indicated that increasing the expression of TRIM35 inhibited number of viable MDA-MB-231 and MCF-7 cells (Fig. 2C and D). The efficacy of TRIM35 overexpression and silencing using multiple clones are provided in Figs. S1 and S2. Furthermore, a Transwell assay suggested that increasing the expression of TRIM35 significantly inhibited the migration of MDA-MB-231 and MCF-7 cells (Fig. 2E and F). In addition, Transwell experiments indicated that increasing the expression of TRIM35 significantly inhibited the invasion of MDA-MB-231 and MCF-7 cells (Fig. 2G and H). Increasing the expression of TRIM35 promoted the apoptosis of MDA-MB-231 and MCF-7 cells (Fig. 2I and J). However, knockdown of TRIM35 promoted the malignant biological behavior of breast cancer cells (Fig. S3).

Metabolic reprogramming of tumor cells is an important biological change, mainly manifested as enhanced glycolysis, which provides effective energy and metabolites for the rapid growth of tumor cells in different microenvironments (25,26).

In the present study, it was analyzed whether TRIM35 affects glycolysis in breast cancer cells. The level of glucose uptake was tested (Fig. 3), revealing that it was significantly decreased after TRIM35-overexpression in MDA-MB-231 and MCF-7 cells (Fig. 3A and E). In addition, the levels of related metabolites, i.e. pyruvate and lactate, as well as the production of ATP, were tested. The results revealed that the levels of pyruvate and lactate, and the production of ATP were all significantly decreased in TRIM35-overexpressing MDA-MB-231 (Fig. 3B-D) and MCF-7 cells (Fig. 3F-H).

The in vivo experiments indicated that overexpression of TRIM35 significantly inhibited the growth of tumors derived from MDA-MB-231 breast cancer cells (Fig. 4A-C). In addition, western blot indicated that overexpression of TRIM35 significantly inhibited the expression of N-cadherin, Twist1, Bcl-2 and Ki-67, while significantly promoting the expression of E-cadherin and Bax in MDA-MB-231-cell tumors (Fig. 4D). In line with the above results, in tumors derived from MCF-7 cells with overexpression of TRIM35, tumor growth was significantly inhibited (Fig. 4E-G) and the expression of N-cadherin, Twist1, Bcl-2 and Ki-67 were significantly inhibited, while the expression levels of E-cadherin and Bax were significantly promoted (Fig. 4H). Immunohistochemical staining also confirmed the above results (Fig. S4).

As a key protein of the Warburg effect, PKM2 is widely involved in the glycolysis process of tumor cells (27,28). The above-mentioned results of the present study indicated that TRIM35 significantly regulates the glycolysis process in breast cancer. Therefore, it was further examined whether TRIM35 has a regulatory effect on PKM2.

The relationships between TRIM35 and PKM2 were verified by using reciprocal Co-IP experiments carried out in MDA-MB-231 and MCF-7 cells. The results confirmed the interaction between endogenous TRIM35 and PKM2 (Fig. 5A and B). Subsequently, a protein-protein interaction network prediction was performed, suggesting that TRIM35 had a significant binding regulatory effect on PKM2 (Fig. S5). However, no changes in the mRNA or protein levels of PKM2 were observed after changing the expression of TRIM35 (Fig. 5C-F).

PKM2 functions to convert phosphoenolpyruvate (PEP) to pyruvate in the final step of glycolysis (29). Three dynamic types of PKM2 exist in mammalian cells, namely monomeric, dimeric and tetrameric PKM2s, but only dimeric and tetrameric PKM2s have pyruvate kinase activity (30). Monomeric PKM2 is in an inactive state. Dimeric PKM2 drives glucose-derived carbon towards glycolysis (30). Tetrameric PKM2 promotes the transfer of glucose-derived carbons in the direction of the respiratory chain (31). The above-mentioned results of the present study indicated that glycolytic function was inhibited after overexpression of TRIM35, which also suggested that the changes in tetramers and dimers of PKM2 may be the key target of TRIM35 in regulating PKM2.

Since TRIM35 is an ubiquitin E3 ligase, it was further hypothesized that TRIM35 regulates the expression of PKM2 at the translational or post-translational level (32). Specifically, in the present study, it was investigated whether TRIM35 affects the tetramer and dimer changes of PKM2 by mediating changes in ubiquitination.

First, dimers of PKM2 were indicated to be highly expressed in breast cancer cell lines (Fig. S6A-D). The ubiquitination levels of PKM2 were enhanced in both the MDA-MB-231 and MCF-7 cell lines with TRIM35 overexpression, which suggests that TRIM35 overexpression allowed more PKM2 to enter the degradation process (Fig. 6A and B). In addition, the change in ubiquitination has a regulatory role in the change in tetramers and dimers of PKM2. Therefore, the effect of the change of ubiquitination on the expression of PKM2 tetramer and dimer was further analyzed. Overexpression of TRIM35 significantly increased the expression of PKM2 tetramer, decreased the expression of PKM2 dimer and increased the ratio of PKM2 tetramer in MDA-MB-231 cells (Fig. 6C-F). The same results were also obtained in MCF-7 cells (Fig. 6G-J). Furthermore, knockdown of TRIM35 significantly decreased the expression of PKM2 tetramer, increased the expression of PKM2 dimer and decreased the ratio of PKM2 tetramer in MDA-MB-231 and MCF-7 cells (Fig. S7A-H). When PKM2 dimer was reduced, PKM2 translocated from the nucleus to the cytoplasm (Fig. S8A-D). The PKM2 activator TEPP-46 has been reported to reduce dimer formation and promote tetramer formation (33). Of note, treatment of breast cancer cells with TEPP-46 significantly inhibited number of viable cells (33), migration and invasion and promoted cell apoptosis (Fig. S9). Collectively, these data suggest that TRIM35 affects the expression of the PKM2 tetramer and dimer by regulating the ubiquitination of PKM2 and regulates the Warburg effect to promote the malignant biological behavior of breast cancer (Fig. 6K).