Bioinformatic data analysis

The pan-cancer dataset (TCGA TARGET GTEx (PANCAN, N = 19,131, G = 60,499) that integrated data from the Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression Project (GTEx) were downloaded from the UCSC database (https://xenabrowser.net/) [14]. Then, UCHL1, DNMT1, DNMT3A/B, and TET1/2/3 expression in different PAM50 subtypes of breast cancers (primary tumors), and corresponding normal tissues (adjacent normal from TCGA and normal tissues from healthy controls from GTEx were compared. The expression values have been further transformed using log2(TPM + 0.001). DNA methylation data (methylation 450k) from this dataset was also extracted. The correlations between gene expression and the β values of the CpGs were assessed by calculating the Pearson’s correlation coefficients.

One recent publication that provided a single-cell and spatially resolved transcriptomics analysis of human breast cancers was analyzed using the Single Cell Portal (https://portals.broadinstitute.org/single_cell) [3].

Cell culture and treatment

MCF-7 cells (representing Luminal A), BT474 cells (representing Luminal B), SUM149, and MDA-MB-468 cells (representing Basal-like and TNBC) were procured from Procell (Wuhan, Hubei, China). MCF-7 cells were grown in Minimum Essential Medium containing NEAA, 10 µg/mL Insulin, 10% fetal bovine serum (FBS), and 1% P/S (100 U/mL Penicilium and 100 µg/mL Streptomycin). BT474 cells were grown in RPMI-1640 medium supplemented with 10 µg/ml Insulin, 2mM L-glutamine, 20% FBS, and 1% P/S. SUM149 and MDA-MB-468 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) with 10% FBS and 1% P/S. All cells were maintained in a cell incubator with 5% CO2 at 37℃.

5-Aza-2’-deoxycytidine (5-Aza-dc), cycloheximide (CHX), MG132 Tamoxifen and Fulvestrant were purchased from MedChemExpress (Monmouth Junction, NJ, USA). To induce hypomethylation, cells were treated with 5-Aza-dc (1 µM) for 48 h.

Lentiviral shRNAs were constructed using the pLKO.1-puro plasmids. The following shRNA sequences were used: shUCHL1#1: 5’-CGGGTAGATGACAAGGTGAAT-3’; sh UCHL1#2: 5’-CCAGCATGAGAACTTCAGGAA-3’; shTET1#1: 5’-CCTCCAGTCTTAATAAGGTTA-3’; shTET1#2: 5’- CCCAGAAGATTTAGAATTGAT-3’; shTET1#3: 5’- GCAGCTAATGAAGGTCCAGAA-3’; shTET2#1: 5’-GCGTTTATCCAGAATTAGCAA-3’; shTET2#2: 5’-CCTCAAGCATAACCCACCAAT-3’; shTET2#3: 5’-GCCAAGTCATTATTTGACCAT-3’; shTET3#1: 5’-ACTCCTACCACTCCTACTATG-3’; shTET3#2: 5’-GCCGAAGCTGTGTCCTCTTAT-3’; shTET3#3: 5’-GAACCTTCTCTTGCGCTATTT-3’; scramble control: 5’- CCTAAGGTTAAGTCGCCCTCG-3’. Wide-type lentiviral UCHL1 (NM_004181) overexpressing particles and the mutant (C90S), Myc-tagged UCHL1 (Myc-UCHL1), C-terminal 3xFlag tagged full-length KLF5 (NM_001730) and the truncated fragments (encoding aa1-312 or aa313-457) were generated using pLV-Puro or pLVX-IRES-Puro-3xFlag.

The pMD2.G and psPAX2 vectors were co-transfected with plasmid DNA into HEK 293T cells using TurboFect (Thermo Fisher Scientific Waltham, MA) at a ratio of 5:2:3 to produce lentivirus. After 72 h, the supernatant was collected, filtered through a 0.45 μm filter, and then centrifuged at 25,000 rpm at 4 °C for 1.5 h. The supernatant was then discarded, and the virus pellet was resuspended in an appropriate virus storage solution and left overnight at 4 °C. The packaged virus was collected, and virus titering and specificity testing were performed. The packaged virus was stored at -80℃ until use. Cells were exposed to lentiviral infection at a multiplicity of infection (MOI) of 10, with 6 µg/ml polybrene present.

Immunofluorescent staining

Immunofluorescent staining was performed following a standard protocol [15]. MDA-MB-468 and SUM149 cells were grown on coverslips. After fixation, permeabilization and blocking, cells were incubated with rabbit anti-KLF5 (1:250, 21017-1-AP, Proteintech, Wuhan, China) and mouse anti-UCHL1 (1:250, 66230-1-Ig, Proteintech). Alexa Fluor 488 labeled secondary anti-rabbit IgG (1: 1000, srbAF488-1, Proteintech) and Alexa Fluor 647 labeled secondary anti-mouse IgG (1: 1000, sms2bAF647-1, Proteintech) were used. Nuclei were counterstained using 4,6-Diamidino-2-phenylindole (DAPI). Immunofluorescent images were captured using IX83 (Tokyo, Japan.

Co-immunoprecipitation

MDA-MB-468 and SUM149 cells were plated in 10-cm plates for endogenous UCHL1 and KLF5 protein interaction analysis and ubiquitination studies. For UCHL1 and KLF5 protein interaction analysis, cell samples were lysed using cell lysis buffer for Western and IP (P0013, Beyotime, Shanghai, China). For ubiquitination studies, cell samples were lysed using RIPA lysis buffer (P0013B, Beyotime), which contains 0.1%SDS and 1% sodium deoxycholate (two robust anionic detergents) that disrupt the non-covalent interactions holding protein complexes together. These ingredients and formation are appropriate co-immunoprecipitation assays for ubiquitin detection according to previous publications [16,17,18]. Endogenous proteins were immunoprecipitated with rabbit anti-UCHL1 (21017-1-AP, Proteintech), anti-KLF5 (14730-1-AP, Proteintech) or anti-EGFR (18986-1-AP, Proteintech) antibody respectively, for 120 h at 4.0 °C. Rabbit IgG serves as the negative control. For exogenous FLAG tagged proteins, immunoprecipitation was performed using rabbit anti-Flag (80010-1-RR, Proteintech). 100 µl Protein A Sepharose bead slurry were added to the sample to capture the immunocomplex for 4 h at 4 °C with gentle agitation. Then, the samples were centrifuged at 220G for 30 s at 4 °C to discard the supernatant. After washing with 0.2%TBST, the immunocomplex was eluted for western blotting assays. Mouse derived primary antibodies were applied for subsequent detection to avoid IgG bands.

Quantitative reverse transcription PCR (RT-qPCR)

RT-qPCR was performed following the method described [19]. In brief, total RNA was isolated from cell, reversely transcribed into cDNA and utilized for template for qPCR. Relative gene expression was normalized with GAPDH and calculated using the 2−ΔΔCT method. Amplification was carried out using the following primers: FGFBP1, forward: 5’-TGGCAAACCAGAGGAAGACTGC-3’; reverse: 5’-GGAACCCGTTCTCTTTTGACCTC-3’; EGFR, forward: 5’-AACACCCTGGTCTGGAAGTACG-3’; reverse: 5’-TCGTTGGACAGCCTTCAAGACC-3’; ESR1, forward: 5’- GCTTACTGACCAACCTGGCAGA-3’; reverse: 5’- GGATCTCTAGCCAGGCACATTC-3’; TET1, forward: 5’-CAGGACCAAGTGTTGCTGCTGT-3’; reverse: 5’- GACACCCATGAGAGCTTTTCCC-3’; UCHL1, forward: 5’- CAGTTCAGAGGACACCCTGCTG-3’; reverse: 5’- CCACAGAGCATTAGGCTGCCTT-3’; GAPDH, forward: 5’-GTCTCCTCTGACTTCAACAGCG-3’; reverse: 5’-ACCACCCTGTTGCTGTAGCCAA-3’.

Western blotting assays

Conventional western blotting was performed following a standard protocol [20]. The following antibodies and dilutions were applied: anti-UCHL1 (1: 1000, 14730-1-AP/66230-1-Ig, Proteintech), anti-KLF5 (1:1000, 66850-1-Ig/21017-1-AP, Proteintech), anti-Ubiquitin (anti-Ub, 1:1000, 10201-2-AP, Proteintech), anti-Flag tag (1:1000, 66008-4-Ig, Proteintech), anti-Myc tag (1:1000, 60003-2-Ig, Proteintech), anti-TET1 (1:1000, 61,443, Proteintech, Wuhan, China), anti-TET3 (1: 1000, ABE290, Merck, Darmstadt, Germany) and anti-β-actin (1:2000, 20536-1-AP, Proteintech). Then, the protein bands were visualized using Enhanced chemiluminescence (ECL) (BeyoECL Star, Beyotime, Shanghai, China).

Gene set enrichment analysis (GSEA)

For GSEA, we obtained the GSEA software (version 3.0) from the GSEA website [21]. Primary basal-like tumor cases in TCGA-BRCA were divided into high expression ( > = 50%) and low expression groups (< 50%) based on the expression levels of UCHL1 or KLF5. The h.all.v7.4.symbols.gmt gene set collection from the Molecular Signatures Database [22] was to evaluate the related pathways and molecular mechanisms. The gene expression profiles and phenotype groupings were set with a minimum gene set size of 5 and a maximum of 5000, with 1000 permutations. Gene sets with p-values of < 0.05 were considered statistically significant.

Bisulfite sequencing PCR (BSP)

BSP assay was performed following a protocol introduced previously [23]. To summarize, the cell samples were subjected to genomic DNA extraction and then treated with sodium bisulfite using the EZ DNA Methylation-Gold kit (Zymo Research, Irvine, CA, USA). The converted DNA was utilized for PCR assays with bisulfite-specific primers (forward: 5’-TAAAATTAAAGATTTTATTAAAAGGATTGT-3’; reverse: 5’-AAAAAAAACAAAAACAAAACCAAAC-3’), which included 15 CpGs in the PCR products. Afterwards, the QIAquick PCR purification kit from QIAGEN in Germany was used to purify the PCR products, which were then cloned into the pGEM-T Easy Vector from Promega in Madison, WI, USA. Subsequently, five bacterial clones that contained the insert were chosen for sequencing. The position of the primers was marked in Supplementary Fig. 1.

Chromatin immunoprecipitation (ChIP)-qPCR

ChIP was performed with a commercial ChIP Assay Kit (Beyotime) [24]. MDA-MB-468 and SUM149 cells with or without lentivirus-mediated TET1 or TET3 knockdown were collected and lysed. Immunoprecipitation was performed using anti-TET1 (61443, Proteintech), anti-TET3 (ABE290, Merck) or rabbit IgG (negative control). Then, the immunoprecipitated chromatin samples were purified and used for qPCR assays. The following primers were used: F:5’-ACCGGCGAGTGAGACTG-3’ and R:5’-CACTGTGAGGCCTGTGC-3′. The position of the primers was marked in Supplementary Fig. 1.

Colony formation

MDA-MB-468 cells with TET1 or TET3 knockdown alone or combined with UCHL1 overexpression were placed into 24-well plates (500 cells per well). Cells were cultured for 14 days with or without Tamoxifen (10 µM) or Fulvestrant (500 nM) treatment. Then, the colonies were fixed, stained, and counted.

Animal studies

Animal studies were conducted following a protocol introduced previously [19]. The animal study was carried out at Jinruijie Biotechnology Service Center in Chengdu, China, and received approval from the institution’s ethics committee (Approval no. 2023056SPPH). All procedures involving animals adhered to the Guide for the Care and Use of Laboratory Animals [25]. Female athymic nude mice (BALB/c-nu), weighing approximately 18–20 g and aged 5–6 weeks, were procured from Vital River Laboratory Animal Technology (Beijing, China). These mice were raised in a specific-pathogen-free (SPF) environment. To initiate the experiment, 5 × 106 MDA-MB-468 cells (with indicating treatment) in a mixture of 0.2 ml PBS and Matrigel matrix (BD Biosciences) in a 1:1 volume ratio were injected subcutaneously into the fourth mammary fat pad. Tumor volume was calculated as length × width2× (π/6). After the tumors became palpable, the mice were randomly assigned to one of four groups (n = 6/group). Once the mean diameter of the tumors reached 5–6 mm, the mice were administered the specified treatment (1 mg per dose via oral gavage daily or vehicle control) for 18 days.

The tumor sizes and body weights of mice were measured every other day until euthanization (on the same day using CO2 asphyxiation). Before euthanization, to obtain the peripheral blood sample, a heparin blood collection tube was used, and the supernatant was collected after centrifugation. The collected peripheral blood sample was then analyzed using an automatic biochemical analyzer (Mindray, China) to determine the concentrations of ALT, AST, creatinine, and urea. Then, the xenograft tumors were excised and utilized for IHC staining of Ki-67 and ERα.

Statistical analysis

GraphPad 8.01 was used to collect and integrate the statistical results. The quantitative data was presented as mean ± SD. To compare two groups, an unpaired T-test was used. Pearson’s r or Spearman’s rho were utilized to assess correlation. To evaluate the differences in multiple groups, one-way ANOVA with post-hoc Tukey’s multiple comparisons test was performed. Significance was set at p < 0.05.