Deubiquitinase UCH-L1 confers paclitaxel resistance via stabilizing PKM2 to promote glycolysis in triple-negative breast cancer
Xisha Chen, Xiaoming Zhou, Yingcai Meng, Ying Zhou, Wenjie Zhang, Liyang Yin, Yingying Shen, Jing Zhong, Taolan Zhang, Xuyu Zu
TL;DR
This study shows that UCH-L1 helps breast cancer cells resist chemotherapy by boosting sugar metabolism, suggesting new treatment strategies.
Contribution
The novel finding is that UCH-L1 stabilizes PKM2 via deubiquitination, promoting glycolysis and chemoresistance in TNBC.
Findings
UCH-L1 overexpression correlates with poor prognosis and chemotherapy resistance in TNBC.
UCH-L1 stabilizes PKM2 by removing K48-linked ubiquitination, enhancing glycolysis.
Inhibiting the UCH-L1/PKM2 axis increases sensitivity to paclitaxel in resistant TNBC cells.
Abstract
Resistance to paclitaxel-based chemotherapy represents a major clinic challenge in triple-negative breast cancer (TNBC). Insights on the regulation genes of chemoresistance and underlying mechanisms in TNBC are waiting for in-depth investigation to address the current treatment bottlenecks. In this study, we identified that ubiquitin carboxyl terminal hydrolase-L1 (UCH-L1) was preferentially overexpressed in TNBC and correlated with worse prognosis as well as poor response to chemotherapy. Upregulation of UCH-L1 attenuated the inhibitory effect of paclitaxel on tumor cells through modulating the aerobic glycolysis, while knockdown of UCH-L1 increased the responsiveness of TNBC cells to the drug both in vitro and in vivo. Coimmunoprecipitation results revealed that the N terminal of UCH-L1 interacts with the C-terminal domain of pyruvate kinase M2 (PKM2). UCH-L1 stabilized PKM2 via…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —the Natural Science Foundation of Hunan Province of China (2024JJ5353) and the Health Research Project of Hunan Provincial Health Commission (W20243132).
- —the National Natural Science Foundation of China (82473223)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsCancer, Hypoxia, and Metabolism · Ubiquitin and proteasome pathways · Ferroptosis and cancer prognosis
Introduction
Breast cancer has ranks first in newly diagnosed cases and remains the leading cause of cancer-related death in females worldwide [1]. Triple-negative breast cancer (TNBC), pathologically classified by the absence of the expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), accounts for about 15–20% of all breast cancers [2]. TNBC is the most challenging subgroup of breast cancers with a high risk of distant metastasis and disease recurrence [3]. Given the lack of validated therapeutic targets and effective targeted therapy, chemotherapeutic drugs like anthracyclines, capecitabine and paclitaxel are clinically used as the first-line treatment of TNBC [4]. Although a proportion of TNBC patients respond well to chemotherapy, a considerable number of patients inevitably suffer from drug resistance and tumor recurrence, leading to limited clinical outcomes and poor prognosis [5]. Hence, identifying the underlying molecules and mechanisms driving TNBC progression and chemoresistance will help develop novel and more effective therapies to enhance the efficacy of standard chemotherapy regimens.
Highly proliferative cancer cells preferentially metabolize glucose via aerobic glycolysis even in the case with sufficient oxygen, and such a phenomenon is also named the Warburg effect [6]. High rate of aerobic glycolysis is recognized as a hallmark of cancers and participates in cancer progression and drug resistance by providing energy and metabolic intermediates [7]. Importantly, aerobic glycolysis is observed upregulated in most chemotherapy-resistant cancer cells, and accumulating evidence suggests that blockade of glycolysis is an effective approach to restoring therapeutic sensitivity [8–10].
Pyruvate kinase M2 (PKM2), a critical regulator of the final rate-limiting step of glycolysis, is upregulated in various cancer cells and facilitates tumor development and metastasis [11]. Post posttranslational modifications (PTMs) of PKM2 including phosphorylation, acetylation, O-GlcNAcylation and ubiquitination affect its enzymatic activity, subcellular localization and protein stability [12–16]. However, deubiquitinating enzymes that promote PKM2 stabilization are far of well recognized currently.
Ubiquitin carboxyl terminal hydrolase-L1 (UCH-L1) is a deubiquitinase that belongs to ubiquitin c-terminal hydrolases family [17]. While ubiquitous in brain and was initially considered to modulate neurodegenerative disorders such as Parkinson’s and Alzheimer’s disease [18], increasing studies have reported that UCH-L1 acts as an oncogene and overexpressed UCH-L1 is associated with aggressive phenotypes and poor prognosis [19–25]. We have previously identified that UCH-L1 inhibits ERα expression and contributes to insensitive of breast cancer cells to endocrine therapy via deubiquitinating and stabilizing EGFR, suggesting UCH-L1 represents a therapeutic target for hormone therapy-insensitive breast cancers [26]. Nevertheless, the roles and mechanisms of UCH-L1 in breast cancer chemoresistance remain largely unknown.
In this study, we show that UCH-L1 is preferentially overexpressed in TNBC and confers paclitaxel resistance. Mechanistically, UCH-L1 binds to PKM2 to deubiquitinate and stabilize PKM2, and thereby promoting glycolysis. Therapeutically targeting UCH-L1-PKM2 axis-mediated glycolysis restores the sensitivity of paclitaxel-resistant TNBC cells to paclitaxel both in vitro and in vivo. Our results provide an understanding of the role of UCH-L1 in breast cancer and facilitate the development of novel treatment strategies to overcome chemotherapy resistance.
Materials and methods
Cell culture
The human breast cancer cell lines HCC1806, BT549 and HCC1937 were cultured in RPMI-1640 medium, MDA-MB-231, MDA-MB-436, MDA-MB-468, Hs578t and human embryonic kidney 293 T (HEK293T) cell lines were cultured in DMEM medium. All the above cell culture media were supplemented with 10% fetal bovine serum, 100 units/mL penicillin and 100 μg/mL streptomycin. The normal human mammary epithelial cell line MCF-10A was cultured in DMEM/Ham’s F12 supplemented with 5% horse serum (Gibco), hydrocortisone, insulin, epidermal growth factor (EGF), and 1% penicillin-streptomycin. All cell lines were maintained in a 5% CO_2_ incubator at 37 °C. Cell lines were authenticated using STR profile analysis and periodically checked to confirm they were mycoplasma-free. All cell lines were used within 3 to 15 passages of thawing the original stocks.
Reagents and antibodies
Paclitaxel (S1150), the proteasome inhibitor MG132 (S2619) and protein synthesis inhibitor cycloheximide (CHX, S7418) were purchased from Selleck (Shanghai, China). Antibodies used in immunoblotting: UCH-L1 (66230-1-Ig, 1:1000), PKM2 (60268-1-Ig, 1:1000), β-actin (66009-1-Ig, 1:5000), Myc (60003-2-Ig, 1:5000), HA (66006-2-Ig, 1:5000), Flag (66008-4-lg, 1:5000), GST (10000-0-AP, 1:4000) were purchased from Proteintech. Antibodies against UCH-L1 (13179, 1:200 for immunofluorescence and IHC staining) were purchased from Cell Signaling Technology. The Peroxidase AffiniPure Goat Anti-Rabbit/Mouse IgG (H + L) was purchased from Jackson Immuno Research.
siRNA, short hairpin RNA, and plasmid transfection
siRNAs targeting UCH-L1 and PKM2 were synthesized by RiboBio (Guangzhou, China). Transfection of siRNA was carried out according to the manufacturer’s protocol. To generate UCH-L1 stable knockdown (KD) cell pools, shRNA sequences targeting UCH-L1 were cloned into GV248-puro lentiviral vectors. The lentiviral particles were produced by co-transfection of pHelper 1.0, pHelper 2.0 and GV248 vector plasmids into 293T cells. TNBC cells were then infected with the lentivirus and selected with puromycin (1 μg/mL) for 1 week. The efficiency of UCH-L1 knockdown in TNBC cells was monitored by Western blot assay.
Using full-length UCH-L1 and PKM2 amplicons as templates, a series of UCH-L1 truncation plasmids (UCH-L1 amino acids 1-100 and 101-223) were amplified by PCR and cloned into the Myc-pcDNA 3.0(+), PKM2-truncated fragments (PKM2 amino acids 1-388, 44-388 and 389-531) were cloned into Flag-tagged destination vectors. The point mutations of PKM2 (K475R, K498R) were generated using a Fast Site-directed Mutagenesis Kit (TransGen Biotech, FM111). The HA-K6, -K11, -K27, -K29, -K33, -K48, -K63, and -Ub plasmids were originally provided by Dr. Yongguang Tao (Central South University). All transfection experiments were carried out using Lipofectamine 3000 reagents (Invitrogen) according to the manufacturer’s protocols. pLVX-IRES-NEO lentiviral vectors were used to stably express pLVX-Flag-UCH-L1. TNBC cells were cultured with lentivirus for UCH-L1 transfection and the antibiotic-resistant transfected cells were selected and enriched by treatment of puromycin at a final concentration of 1 μg/ml in culture medium. The efficiency of overexpressed-UCH-L1 in TNBC cells was evaluated by Western blot assay.
Immunoblotting (Western blot) analysis
Protein lysates from cultured cells were lysed with cold RIPA lysis buffer supplemented with a protease inhibitor cocktail (Bimake, B14002). After measuring the protein concentrations by BCA protein assay, 20–40 μg of total protein was loaded and separated by SDS-PAGE and then transferred to PVDF membrane. The PVDF membranes were incubated with the indicated antibodies in 5% BSA at 4 °C overnight, followed by incubation with a secondary antibody at room temperature for 1 h. The protein signals were then visualized by ECL method.
Immunoprecipitation assays and proteomics analysis
For immunoprecipitation (IP), NP-40 buffer containing protease inhibitor cocktail was used for cell lysis. Immunoprecipitations were performed using the indicated primary antibody and protein A/G plus agarose (Santa Cruz, sc-2003) at 4 °C overnight. The immunocomplexes were then washed four times with NP-40 buffer, and proteins were boiled in SDS–PAGE sample buffer for 10 min, followed by immunoblotting analysis. IP coupled with LC-MS/MS analysis was performed to examine UCH-L1-interacting proteins. 4D-DIA quantitative proteomic analysis was performed to identify downstream targets of UCH-L1 (Wininnovate Bio, Shenzhen, China).
Protein stability assay
For the detection of the protein half-life of PKM2, cells were treated with CHX (50 μg/mL) for the indicated time periods (0, 4, 8, and 12 h) before collection, followed by Western blot analysis.
GST pulldown assay
Recombinant GST-UCH-L1 protein or GST protein bound to glutathione-sepharose 4B beads was incubated with cell lysates at 4 °C for 4 h. Then the beads were washed with GST binding buffer four times, and the bound proteins were separated by SDS-PAGE, followed by immunoblotting analysis with the indicated antibodies.
In vivo deubiquitination assay
In vivo deubiquitination assay was performed in HEK293T and TNBC cells. HEK293T cells were transfected with HA-Ub, Flag-PKM2, Myc-UCH-L1 or Myc-UCH-L1^C90S^ plasmid as indicated for 48 h. In TNBC cells, HA-Ub and UCH-L1 siRNA or UCH-L1 plasmids were transfected into HCC1806, BT549 or MDA-MB-231 cell lines. After treatment with 20 μM of the proteasome inhibitor MG132 for 8 h, cells were lysed and immunoprecipitated with anti-Flag antibody and immunoblotted with anti-HA.
In vitro deubiquitination assay
HEK293T cells were transfected with HA-Ub, Flag-PKM2 or empty vector plasmid. Two days after transfection, the cells were treated with 20 μM MG132 for 8 h to enrich the ubiquitinated PKM2 proteins. Then, Flag-tagged ubiquitinated PKM2 proteins were purified by immunoprecipitation and eluted with Flag peptides (Sigma). Next, the ubiquitinated PKM2 proteins were incubated with bacterially purified GST or GST-UCH-L1 (wild-type or C90S) proteins in deubiquitination buffer (50 mM Tris-HCl, 50 mM NaCl, 1 mM EDTA, 10 mM DTT, 5% glycerol, pH 8.0) at 37 °C for 2 h, followed by the addition of SDS to a final concentration of 2%. The reaction mix was then boiled at 95 °C, diluted 10–20 times via the deubiquitination buffer, and then immunoprecipitated by Flag beads. The proteins were released from the beads by boiling in SDS–PAGE loading buffer and analyzed by immunoblotting with anti-HA antibody.
Cell proliferation analysis
For the cell viability assay, 5 × 10^3^ TNBC cells were seeded into 96-well microplates and subjected to different treatments after cell adhesion. CCK8 solution reagent was added to each well and incubated for 1–2 h at 37 °C, then the absorbance of the colored solution was detected with a spectrophotometer at 450 nm wavelength.
The EdU incorporation assay was conducted as we previously reported [26]. Briefly, cells were incubated with 5-Ethynyl-2’-deoxyuridine assay (EdU; Ribobio) for 2 h, and processed according to the manufacturer’s instruction. After three washes with PBS, the cells were incubated with 100 μL of 1X Apollo reaction cocktail for 30 min. Then cells were washed three times with 0.5% Triton X-100. The DNA contents were stained with 100 μL of 1X Hoechst 33342 (5 μg/mL) for 30 min and visualized under a fluorescence microscope.
For the colony formation assay, TNBC cells were seeded into 6-well plates or 35 mm cell culture dishes. After 12–15 days incubation, cells were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet for 20 min, washed with PBS, and then the colonies were counted.
TUNEL assay
For the conduction of apoptosis analyses, cells subjected to the indicated treatment were stained with Terminal-deoxynucleoitidyl Transferase Mediated Nick End Labeling (TUNEL, Abbkine, KTA2011) for 60 min at 37 °C, and then washed three times with PBS. Nuclei were stained with DAPI for 10 min at room temperature.
Measurement of the levels of aerobic glycolysis
The TNBC cells were seeded at a density of 1.0 × 10^5^ cells/well in a 6-well plate. Culture media were collected at 48 h after treatment, and the Glucose Kit and Lactate Assay Kit (Nanjing Jiancheng Bioengineering Institute, China) were used to detect the concentrations of glucose and lactate in the medium. At the same time, cells were collected and ATP production was measured with an ATP assay kit according to the manufacturer’s protocol.
Real-time quantitative PCR
Total RNAs were isolated from cells using the Trizol reagent and 1st strand cDNA was synthesized using PrimeScript RT Reagent Kit (Perfect real-time) (Takara). Real-time PCR was carried out using the SYBR green mix with the CFX96TM Real-time PCR Detection System (Bio-Rad, USA). The 2^−ΔΔCt^ method was used for quantification of the relative gene expression. β-actin expression was used for normalization.
Measurement of extracellular acidification rate (ECAR)
The ECAR of TNBC cells was detected by the Seahorse XFe Extracellular Flux Analyzer. Briefly, HCC1806 cells transfected with UCH-L1 siRNAs or control siRNA were seeded in an XFe plate and cultured overnight. The next day, cells were incubated with fresh assay medium after washing and sequentially exposed to 10 mM glucose, 1 μM ATP synthase inhibitor oligomycin and 50 mM 2-DG at the indicated time points. The data were measured by the Seahorse software and normalized to cell number.
Immunofluorescence (IF) and immunohistochemistry (IHC) assay
HCC1806 cells seeded on glass coverslip were fixed in 4% paraformaldehyde for 20 min at room temperature and blocked in 5% bovine serum albumin (BSA) for 1 h. Then, cells were incubated with anti-UCH-L1 antibody and anti-PKM2 antibody at 4 °C overnight, followed by incubation with Alexa Fluor 594 dye-conjugated anti-mouse IgG antibody and Alexa Fluor 488 dye-conjugated anti-rabbit IgG antibody at room temperature for 1 h. At the end of incubation, the cells were stained with 4,6-diamidino-2-phenylindole (DAPI). The coverslip was washed in phosphate-buffered saline (PBS) and fluorescent signals were captured using a confocal microscope.
IHC staining for UCH-L1, PKM2 and Ki67 was performed on formalin-fixed paraffin-embedded tissues using the DAKO LSAB+System-HRP kit (DAKO), according to the manufacturer’s instructions. IHC staining was assessed by two independent pathologists under blinded conditions. UCH-L1 and PKM2 expression were scored by the staining intensity and the percentage of positively stained immunoreactive cells. The intensity was rated as follows: “0” points no brown particle staining, “1” for light brown particles, “2” for moderate brown particles, and “3” mean dark brown particles. The percentage of positive cells was also classified into four scores: “0” points <10% positive cells, “1” for 10-40% positive cells, “2” for 40–70% positive cells, and “3” for ≥70% positive cells. The two scores were multiplied and used to determine low (score <3) or high (score ≥3) expression of UCH-L1 and PKM2.
Databases description
GEPIA 2 (Gene Expression Profiling Interactive Analysis 2, http://gepia2.cancer-pku.cn/) was used to compare UCH-L1 and PKM2 expressions between tumor and normal tissues of breast cancer samples. Kaplan–Meier analysis of UCH-L1 and PKM2 in TNBC was performed in Kaplan–Meier Plotter (http://www.kmplot.com). ROC plotter server (https://rocplot.org/) was utilized to analyze the correlation between the expression of UCH-L1, PKM2 and the therapeutic responses of breast cancer patients to chemotherapy. Ubiquitination sites prediction of PKM2 was conducted in GPS-Uber (http://gpsuber.biocuckoo.cn/) and GeneCards (https://www.genecards.org/). Gene expression from GSE90564 was extracted from the NCBI Gene Expression Omnibus (GEO) database for analysis of UCH-L1 and PKM2 expressions in paclitaxel-resistant and parental cells.
Animal studies
Animal studies were approved by the Animal Ethics Committee of the University of South China (2023uscxs180). Briefly, 2 × 10^6^ TNBC cells were subcutaneously injected into 6-week-old female nude mice. Tumor sizes were measured as indicated, and calculated as tumor volumes = length × width^2^ × (π/6). Seven days post cell injection, paclitaxel (10 mg/kg, dissolved in DMSO and diluted with PEG300 and Tween80) or vehicle was administered by intraperitoneal injection every 3 days and 6 times in total. At the end of the experiment, mice were anaesthetized and then euthanized by cervical dislocation, and tumors were sectioned and formalin-fixed paraffin-embedded slides were made for IHC staining.
Statistical analysis
The results are presented as mean ± SD from at least three independent experiments. Comparison between the two groups was analyzed using Student’s t-test. Comparison of multiple groups (>2) was performed using one-way ANOVA. The Kaplan-Meier survival curve and log-rank test were conducted to analyze the survival. Mann-Whitney test was used for the analysis of associations between the response to chemotherapy and the indicated gene expression. Pearson correlation test was conducted to analyze the correlation between UCH-L1 and PKM2 expression in clinical samples. GraphPad was used to perform statistical analysis. P value < 0.05 was considered statistically significant.
Results
UCH-L1 is aberrantly expressed in TNBC and predicts poor prognosis and chemotherapeutic response
We have previously uncovered an important role of UCH-L1 in modulating ERα status and in dictating anti-estrogen therapeutic response in breast cancer. To elucidate the biological functions and clinical significances of UCH-L1 in depth, we further evaluated UCH-L1 expression in different subtypes of breast cancer using the GEPIA database. Compared to other subtypes of breast cancer, UCH-L1 was highly expressed in TNBC (Fig. 1A). More importantly, the mRNA levels of UCH-L1 were significantly higher in TNBC tissues than in normal breast tissues (Fig. 1B). To verify the results from the bioinformatics analyses, we assessed UCH-L1 protein expression in 40 samples of TNBC tissues and adjacent normal tissues by immunohistochemistry (IHC) staining. Quantitative analyses of the IHC staining demonstrated a marked increase in UCH-L1 expression in TNBC (Fig. 1C). Consistently, the abundance of UCH-L1 was elevated in multiple TNBC cell lines than in normal breast cell line (Fig. 1D).Fig. 1UCH-L1 is upregulated in TNBC and predicts poor prognosis and chemotherapeutic response.A Boxplot of the UCH-L1 mRNA expression levels in four different subtypes of BRCA from the GEPIA2 database. B mRNA expression levels of UCH-L1 in TNBC and normal breast tissues from the GEPIA2 database. C IHC analysis of TNBC (n = 40) and normal breast tissue samples (n = 20) was performed using anti-UCH-L1 antibody. Representative images of IHC staining and quantified data presented in box plots are shown. Scale bar, 100 μm. D Western blot analysis of UCH-L1 protein expression level in normal breast cell and multiple TNBC cell lines. β-Actin was used as a loading control. E Kaplan-Meier plots for the overall survival of TNBC patients according to UCH-L1 mRNA expression from the Kaplan–Meier Plotter database. F Kaplan-Meier plots for the relapse-free survival of TNBC patients according to UCH-L1 protein expression from the Kaplan–Meier Plotter database. G Kaplan-Meier plots for the overall survival of chemotherapy-treated TNBC patients according to UCH-L1 mRNA expression from the Kaplan–Meier Plotter database. H The receiver operating characteristic (ROC) curve between UCH-L1 expression and therapeutic responses to taxane-containing NAC in breast cancer cohorts. I The correlation analysis between the efficacy of taxane-based chemotherapy and UCH-L1 expression. Chi-square test was performed to calculate statistical significance between two comparator groups. Stacked bar chart showed the proportion of patients with or without response in high or low UCH-L1 groups. *p < 0.05; ***p < 0.001.
We next validated the prognostic value of UCH-L1 in TNBC and found that a high expression of UCH-L1 indicated a poor prognosis of TNBC patients as evidenced by the Kaplan–Meier overall survival (OS), relapse-free survival (RFS) and distant metastasis-free survival (DMFS) analysis with the Kaplan–Meier Plotter database (Fig. 1E, F, Fig. S1A). Moreover, in a cohort of TNBC patients treated with chemotherapy, patients with high UCH-L1 expression showed a significantly shorter OS compared to patients with low UCH-L1 levels (Fig. 1G); nevertheless, the prognostic value of UCH-L1 mRNA was not significant in those without chemotherapy (Fig. S1B). These results indicated UCH-L1 is of high significance as a potential gene involved in the chemoresistance of TNBC. To confirm this hypothesis, we explored the association of UCH-L1 expression and chemo-therapeutic response based on 5-year relapse-free survival in breast cancer patients via ROC Plotter-Online ROC analysis database, and the results revealed that the non-responder group showed upregulated UCH-L1 expression (Fig. S1C). We further compared the UCH-L1 expression between pre-neoadjuvant chemotherapy (NAC) samples from responder and non-responder groups in a breast cancer cohort that underwent taxane-containing NAC. As shown in Fig. 1H, patients who do not exhibit response to taxane had a significantly higher level of UCH-L1 expression (Fig. 1H). When we divided the patients into high and low UCH-L1 expressing groups, the proportion of patients responding to taxane treatment was significantly lower in high UCH-L1 level group than that in low UCH-L1 level group (Fig. 1I). Similar results were also observed in patients who received anthracycline-based chemotherapy (Fig. S1D, E). To sum up, the above data implied that highly expressed UCH-L1 indicates poor prognosis and chemotherapy response in TNBC patients.
Knockdown of UCH-L1 increases paclitaxel sensitivity of TNBC in vitro and in vivo
Based on the accumulated evidence, we sought to verify whether UCH-L1 regulates the TNBC cell response to paclitaxel (PTX), a commonly used chemotherapy drug in treating breast cancer patients. We stably transfected MDA-MB-231 cells with UCH-L1 plasmid, and confirmed a remarkable increase of UCH-L1 level (Fig. S2A). Overexpression of UCH-L1 caused a notable decrease of sensitivity to paclitaxel (Fig. S2B-D), implying upregulation of UCH-L1 contributes to paclitaxel resistance. On the other hand, we stably silenced UCH-L1 utilizing short hairpins (shRNA) in HCC1806 and BT549 cells, two TNBC cell lines with high level of UCH-L1 expression. Result from western blot assay demonstrated the two shRNA sequences worked well in both cell lines (Fig. 2A). Silencing UCH-L1 dramatically increased PTX-induced decreases in cell viability (Fig. 2B). Consistently, colony-formation assays (Fig. 2C), Edu assays (Fig. 2D) and TUNEL assays (Fig. 2E) revealed that UCH-L1 knockdown sensitizes TNBC cells to paclitaxel treatment.Fig. 2UCH-L1 knockdown sensitizes TNBC cells to paclitaxel treatment.A Knockdown efficiency of UCH-L1 in TNBC cells was evaluated by Western blot assay. B HCC1806 and BT549 cells with or without UCH-L1 depletion were incubated with indicated concentrations of PTX for 72 h, and cell viability was measured by CCK-8 assay. C Colony formation of HCC1806 and BT549 cells (500 cells per well) with or without UCH-L1 depletion in the presence of DMSO or 20 nM PTX. D HCC1806 and BT549 cells with or without UCH-L1 depletion were incubated with 20 nM PTX for 24 h, and cell proliferation was measured by EdU assay. Scale bars, 150μm. E HCC1806 and BT549 cells with or without UCH-L1 depletion were incubated with 20 nM PTX for 48 h, and cell apoptosis was measured by TUNEL assay. Scale bars, 150μm. F A schematic view of the animal study plan (n = 6). G Curves of tumor growth. H Photopraphs of mice tumors of each group at the termination of the experiment. I Tumor weights were measured at the end of the experiment. J Characterization of HCC1806 xenograft tumors with histologic analysis by IHC staining of Ki67 and TUNEL. K The effect of treatment on mice body weight. Data are represented as mean ± SD of biological triplicates. *p < 0.05; **p < 0.01; ***p < 0.001.
To confirm the anti-chemotherapy function of UCH-L1 in vivo, HCC1806 cells with or without UCH-L1 depletion were subcutaneously injected into the nude mice and then subjected to paclitaxel treatment (Fig. 2F). Compared to the control group, administration of paclitaxel suppressed the xenograft tumor growth, and in combination with UCH-L1 depletion further enhanced the inhibitory effect, as evidenced by the greater reductions in tumor volume and tumor weight (Fig. 2G-I). Moreover, the histopathologic characterization of the xenograft tumors demonstrated that Ki67-positive cells were obviously reduced in both shUCH-L1 and PTX groups, and a greater decrease of the ratio of Ki67-positive cells was observed in the group of shUCH-L1 and PTX co-treatment (Fig. 2J); TUNEL staining results of xenograft tumor tissues showed that UCH-L1 knockdown increased PTX-induced apoptosis (Fig. 2J). There was no significant difference in body weight among the groups of mice (Fig. 2K). Taken together, our results suggested knockdown of UCH-L1 significantly increases the sensitivity of TNBC cells to paclitaxel.
UCH-L1 enhances aerobic glycolysis
To gain further insights into the mechanism by which UCH-L1 regulates chemoresistance, we carried out quantitative proteomics experiment and then performed gene set enrichment analysis (GSEA) for pathway enrichment based on differentially expressed genes from HCC1806 cells with UCH-L1 knockdown and control cells (Fig. 3A). In addition to the known functions of regulating neurodegenerative disorders like Alzheimer disease, the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that UCH-L1 expression was dramatically correlated with glycolysis pathway of cancer (Fig. 3B). Aberrant glycolysis has been reported to be responsible for tumor resistance to conventional chemotherapy [27, 28], so we examined the effect of UCH-L1 on glycolysis. In comparison with control HCC1806 cells, UCH-L1-deficient cells displayed markedly decreased extracellular acidification rate (ECAR) (Fig. 3C). A quantitative analysis showed that UCH-L1 knockdown inhibited the overall glycolytic flux including glycolysis, glycolytic capacity as well as glycolytic reserve (Fig. 3D). Simultaneously, UCH-L1 deficiency significantly reduced glucose consumption, lactate production and cellular ATP levels (Fig. 3E-G). In support of this observation, we transfected MDA-MB-231 cells with the Myc-UCH-L1 plasmid and found that excessive expression of UCH-L1 strikingly increased glucose consumption, lactate production and cellular ATP levels (Fig. 3H-J). Collectively, these results indicate UCH-L1 promotes glycolysis in TNBC. To verify whether UCH-L1-regulated glycolysis mediates paclitaxel resistance, we treated UCH-L1-overexpressing MDA-MB-231 cells with the glycolysis inhibitor 2-DG, and observed that the decreased sensitivity to paclitaxel caused by UCH-L1 upregulation was reversed by 2-DG treatment (Fig. 3K-O). The TUNEL assay in HCC1806 cells demonstrated that UCH-L1 overexpression inhibited PTX-induced cell apoptosis, while this effect was rescued by the addition of 2-DG (Fig. S2E), which further indicated 2-DG could block UCH-L1-induced paclitaxel resistance. Therefore, these findings strongly support the point that UCH-L1 promotes chemoresistance due to the enhancement of glycolysis.Fig. 3UCH-L1 enhances the aerobic glycolysis.A Volcano plot of quantitative proteomics showed the differentially expressed genes between UCH-L1 knockdown and control HCC1806 cells. B KEGG pathway analysis for differentially expressed proteins in HCC1806 cells with or without UCH-L1 knockdown. C Effect of UCH-L1 knockdown on the ECAR of HCC1806 cells was detected using the Seahorse analyzer. D Quantification analysis of glycolysis, glycolytic capacity and glycolytic reserve levels in the ECAR assays. E-G Glucose consumption, lactate and ATP production were determined in control and UCH-L1 knockdown HCC1806 and BT549 cells. H-J Glucose consumption, lactate and ATP production were determined in MDA-MB-231 cells with or without UCH-L1 overexpression. K-L Colony formation of MDA-MB-231 cells with or without UCH-L1 overexpression in the presence of DMSO or PTX and 2-DG. M MDA-MB-231 cells from the indicated groups were incubated with PTX for 72 h, and cell viability was measured by CCK-8 assay. N MDA-MB-231 cells stably expressing empty vector (EV) or flag-tagged UCH-L1 were incubated with DMSO or PTX and 2-DG for 24 h, and cell proliferation was measured using EdU. Scale bars, 150 μm. O MDA-MB-231 cells with or without UCH-L1 overexpression were incubated with DMSO or PTX and 2-DG, and cell apoptosis was measured using TUNEL assay. Scale bars, 100 μm. Results shown are representative of three independent experiments. Data are represented as mean ± SD of biological triplicates. *p < 0.05; **p < 0.01; ***p < 0.001.
N-terminal of UCH-L1 interacts with C-terminal of PKM2 to promote chemoresistance
To understand how UCH-L1 regulates glycolysis and thus mediates TNBC chemosensitivity, we performed immunoprecipitation (IP) assay with HCC1806 cell lysates using anti-UCH-L1 antibody, followed by LC-MS/MS analysis to identify UCH-L1-interacting partners. By this approach, a total of 262 proteins that specifically interacted with UCH-L1 were identified according to the cutoff value of identified unique peptide number over two and confidence over 95%. Among the top 20 potential UCH-L1-interacting proteins based on the number of identified unique peptides (Fig. 4A), PKM2 has attracted our attention for its critical role in tumor metabolism. We further co-transfected 293 T cells with Myc-UCH-L1 and Flag-PKM2 plasmids, and identified Flag-PKM2 in Myc-UCH-L1 precipitate but not in Myc-vector precipitate; consistently, Myc-UCH-L1 was presented in anti-Flag co-IPs from cells co-transfected with Myc-UCH-L1 and Flag-PKM2, but not in the cells transfected with Myc-UCH-L1 and Flag-vector (Fig. 4B). The interaction between UCH-L1 and PKM2 was also confirmed by endogenous reciprocal IP assays in BT549 cells (Fig. 4C). Moreover, immunofluorescence staining revealed that UCH-L1 and PKM2 were mainly co-localized in the cytoplasm in HCC1806 cells (Fig. 4D). GST-pull-down experiment indicated UCH-L1 interacted with PKM2 directly in vitro (Fig. 4E).Fig. 4N-terminal of UCH-L1 interacts with C-terminal domain of PKM2 to promote chemoresistance.A The top twenty UCH-L1-interacting proteins according to the number of identified unique peptides by Co-IP-MS assay. B HEK293T cells were transfected with Flag-PKM2 and Myc-UCH-L1 plasmids, and then subjected to immunoprecipitation with anti-Flag or anti-Myc antibodies. The lysates and immunoprecipitates were then blotted. C Immunoprecipitation analysis with the indicated antibodies was performed to detect endogenous UCH-L1 and PKM2 interaction in BT549 cells. D Immunofluorescence assay to confirm the colocalization of UCH-L1 and PKM2 in HCC1806 cells. E Purified recombinant GST-UCH-L1 interacts with PKM2. GST-UCH-L1 and GST proteins were pulled down with glutathione beads. PKM2 was detected by Western blot. F Schematic representation of full-length (FL) UCH-L1, PKM2 and different truncation mutants. G Myc-UCH-L1 FL or indicated truncation mutants were co-expressed with Flag-PKM2 in HEK293T cells. Flag-PKM2 was immunoprecipitated with anti-Flag antibody, followed by immunoblotting analysis of Myc-UCH-L1 using Myc antibody. H Flag-PKM2 FL or indicated truncation mutants were co-expressed with Myc-UCH-L1 in HEK293T cells. Myc-UCH-L1 was immunoprecipitated with anti-Myc antibody, followed by immunoblotting analysis of Flag-PKM2 using Flag antibody. I MDA-MB-231 cells transfected with the indicated plasmid were incubated with PTX for 72 h, and cell viability was measured by CCK-8 assay. J, K MDA-MB-231 cells with UCH-L1 overexpression were transfected with PKM2 siRNA, and cell viability was detected by CCK-8 assay after 72 h incubation of the indicated concentrations of PTX. Results shown are representative of three independent experiments. Data are represented as mean ± SD of biological triplicates. *p < 0.05; **p < 0.01; ***p < 0.001.
To clarify the structural region responsible for the interaction between UCH-L1 and PKM2, we established truncation plasmids of each other based on their structural domains (Fig. 4F). Results from immunoprecipitation assays demonstrated that N-terminal (1-100aa) domain of UCH-L1 physically interacted with the C-terminal (389-531aa) domain of PKM2 (Fig. 4G, H). Importantly, when subjected MDA-MB-231 cells transfected with indicated constructs to paclitaxel, the survival fraction of UCH-L1-CTD group had no significant difference in comparison with its negative control but was significantly lower than that of UCH-L1-WT group; whereas, the survival fraction of UCH-L1-NTD group was equivalent to that of UCH-L1-WT group (Fig. 4I). It was also found that depletion of PKM2 abolished the chemoresistance effect induced by UCH-L1 overexpression (Fig. 4J, K). Taken together, these results imply that UCH-L1 does rely on its interaction with PKM2 to confer paclitaxel resistance.
UCH-L1 stabilizes PKM2 through the deubiquitylation activity
To verify the functional consequences of this interaction, we investigated PKM2 expression and stability upon UCH-L1 knockdown or overexpression as it blocks degradation of targeted proteins via deubiquitination. It was found that depletion of UCH-L1 decreased PKM2 protein levels in HCC1806 and BT549 cells (Fig. 5A). Ectopic expression of UCH-L1 profoundly elevate PKM2 in a dose-dependent manner; while the catalytically inactive mutant C90S (UCH-L1^C90S^) lost its ability to upregulate PKM2, indicating UCH-L1 regulates PKM2 in a DUB activity-dependent manner (Fig. 5B). We then re-expressed wild-type (WT) UCH-L1 and inactive mutant UCH-L1^C90S^ in UCH-L1 knockdown HCC1806 cells and observed that UCH-L1^WT^, but not mutant UCH-L1^C90S^, remarkably upregulated PKM2 protein levels (Fig. 5C). RT-qPCR analysis suggested UCH-L1 did not affect the mRNA abundance of PKM2 (Fig. S3). Furthermore, addition of the proteasome inhibitor MG132 reversed the decrease of PKM2 protein caused by UCH-L1 deletion (Fig. 5D). We next utilized cycloheximide (CHX) chase assay to interrogate the effect of UCH-L1 on the stability of PKM2 protein. Overexpression of UCH-L1^WT^, but not mutant UCH-L1^C90S^, prolonged the half-life of PKM2 (Fig. 5E). In accordance with these results, loss of UCH-L1 decreased PKM2 protein stability (Fig. 5F). Altogether, these results indicate that UCH-L1 stabilizes and maintains PKM2 dependent on its deubiquitylation activity.Fig. 5UCH-L1 stabilizes PKM2 through the deubiquitylation activity.A Immunoblotting analysis of PKM2 and UCH-L1 in UCH-L1 knockdown TNBC cells. B Immunoblotting analysis of PKM2 and UCH-L1 in MDA-MB-231 and HEK293T cells transfected with increasing amounts of UCH-L1 WT or C90S. C HCC1806 cells with UCH-L1 depletion were transfected with UCH-L1 WT or C90S, and PKM2 expression was detected. D Immunoblotting analysis of PKM2 and UCH-L1 in UCH-L1 knockdown TNBC cells treated with or without MG132. E MDA-MB-231 cells transfected with UCH-L1 WT or C90S were treated with CHX for the indicated time. F Immunoblotting analysis of PKM2 and UCH-L1 in UCH-L1 knockdown TNBC cells treated with or without CHX for the indicated time. PKM2 expression was detected by Western blot. Quantitation of PKM2 protein level based on band intensity was shown. Results shown are representative of three independent experiments.
UCH-L1 deubiquitylates PKM2
As a deubiquitylase, UCH-L1 protects substates from degradation via cleaving the ubiquitin chains from targeted proteins. Therefore, we went on to investigate the possibility that UCH-L1 deubiquitylates PKM2. Ectopic expression of UCH-L1^WT^, but not UCH-L1^C90S^, markedly abrogated PKM2 ubiquitylation in 293 T cells (Fig. 6A and Fig. S4A). Similar results were observed in MDA-MB-231 breast cancer cell line (Fig. 6B). Knockdown of UCH-L1 enhanced PKM2 ubiquitylation in both HCC1806 and BT549 cells (Fig. 6C and Fig. S4B). We further determined whether UCH-L1 directly deubiquitylated PKM2 in vitro. Glutathione S-transferase (GST)-UCH-L1 and GST-UCH-L1^C90S^ were expressed and purified from an Escherichia coli system. Polyubiquitinated PKM2 proteins purified from 293T cells by immunoprecipitation were incubated with GST, GST-UCH-L1 or GST-UCH-L1^C90S^. Compared with GST and UCH-L1^C90S^, UCH-L1 specifically decreased PKM2 polyubiquitination (Fig. 6D). We further conducted an ubiquitination assay with a series of ubiquitin mutants to clarify which type of ubiquitin chain of PKM2 was deubiquitylated by UCH-L1. Of note, UCH-L1 appeared to efficiently remove K48-linked polyubiquitin chains from PKM2 (Fig. 6E). Taken together, these findings suggest that UCH-L1 is a specific DUB for PKM2 and that UCH-L1 de-polyubiquitylates PKM2 to stabilize it.Fig. 6UCH-L1 deubiquitylates PKM2.A Flag-PKM2 and HA-ubiquitin (HA-Ub) were co-expressed with Myc-UCH-L1 WT or C90S in HEK293T cells. After MG132 treatment, IP was performed with Flag antibody, followed by immunoblotting with the indicated antibodies. B HA-Ub was co-expressed with Myc-UCH-L1 WT or C90S in MDA-MB-231 cells. After MG132 treatment, IP was performed with PKM2 antibody, followed by immunoblotting with indicated antibodies. C HCC1806 cells with or without UCH-L1 depletion were transfected with HA-Ub. IP was performed with PKM2 antibody after MG132 treatment, followed by immunoblotting with indicated antibodies. D Ubiquitinated PKM2 was purified from HEK293T cells transfected with HA-Ub, and then incubated with GST-tagged wild-type UCH-L1, and UCH-L1 C90S proteins in reaction buffer. The reactions were terminated by boiling in 1X SDS sample buffer and analyzed using Western blot with anti-HA antibody. E HA-WT, K6, K11, K27, K29, K33, K48, or K63 Ub were co-transfected with Flag-PKM2 and Myc-UCH-L1 into HEK293T cells. Cell lysates were collected after MG132 treatment for 8 h and subjected to ubiquitination assay, and the ubiquitination level of PKM2 was detected by HA antibody. F Flag-PKM2 (WT or indicated KR mutants) and HA-Ub were co-expressed in 293T cells with or without UCH-L1 knockdown. IP was performed with Flag antibody after MG132 treatment for 8 h, followed by immunoblotting with anti-HA. G HCC1806 cells expressing Flag-PKM2 WT or indicated KR mutants were transfected with UCH-L1 siRNA, followed by immunoblotting with indicated antibodies. H HCC1806 cells expressing indicated KR mutants of Flag-PKM2 were transfected with UCH-L1 siRNA, and then treated with CHX for the indicated time. Cell lysates were collected and subjected to immunoblotting analysis. Quantitation of Flag-PKM2 protein level based on band intensity was shown. Results shown are representative of three independent experiments.
To find out which lysine is required for UCH-L1-hydrolyzed PKM2 ubiquitination, we predicted the potential ubiquitination sites of PKM2 protein using the online bioinformatic tools (GPS-Uber and GeneCards) (Fig. S5A, B). A total of 10 overlapped sites were identified in the full-length of PKM2 protein (Fig. S5C), including two ubiquitination sites located at the C terminus which were conserved among species (Fig. S5D, E). We then mutated the lysine residues of PKM2 to identify the specific sites that are deubiquitinated by UCH-L1. As shown in Fig. 6F, UCH-L1 knockdown obviously elevated the ubiquitination level of wild-type as well as K475R mutant of PKM2, but not K498R mutant. Moreover, depletion of UCH-L1 led to significant decrease in the protein level of wild-type and K475R mutant of PKM2 but had no apparent effect on the K498R mutant (Fig. 6G). In addition, knocking down UCH-L1 obviously shortened the half-life of PKM2 K475R mutant but had no marked effect on that of the K498R mutant (Fig. 6H). These results indicate that K498 is a major site of PKM2 deubiquitination mediated by UCH-L1.
Inhibition of UCH-L1-PKM2 axis-mediated glycolysis reverses paclitaxel resistance
With glycolysis reported to be involved in chemotherapy resistance, we determined whether inhibition of UCH-L1-PKM2 axis-mediated glycolysis could restore response to paclitaxel in resistant cells. To this end, we first analyzed the expression levels of UCH-L1 and PKM2 in paclitaxel-resistant TNBC samples from GEO database. Based on the RNA-seq dataset from GSE90564, both UCH-L1 and PKM2 mRNA expressions were augmented in paclitaxel-resistant cells when compared with their respective parental cells (Fig. S6A). We also measured UCH-L1 and PKM2 levels in established HCC1806/TaxR (paclitaxel-resistant) cells (Fig. 7A), and in consistence with the above observation, HCC1806/TaxR cells revealed increased UCH-L1 and PKM2 protein expressions (Fig. 7B). In addition, the glycolysis level was also elevated in HCC1806/TaxR cells (Fig. 7C). We then examined the effect of suppression of UCH-L1-PKM2-glycolysis on paclitaxel sensitivity in HCC1806/TaxR cells. Compared with the control cells, cells with UCH-L1 or PKM2 knockdown or treated with glycolysis inhibitor 2-DG all exhibited decreased resistance to paclitaxel treatment (Fig. 7D). The TUNEL results demonstrated PTX treatment alone hardly induced apoptosis of HCC1806/TaxR cells, but when combined with UCH-L1 depletion or 2-DG, the proportions of apoptotic cells were significantly increased (Fig. 7E). Subsequent Edu and colony formation assays confirmed that knockdown of UCH-L1 restored the inhibitory effect of PTX on HCC1806/TaxR cells’ proliferation (Fig. 7F-G). More importantly, the re-sensitization effect of UCH-L1 depletion on paclitaxel could be largely blocked by PKM2 reconstitution (Fig. S6B-E). To confirm the role of UCH-L1-PKM2 axis in paclitaxel resistance in vivo, we generated a xenograft nude mouse model. It was found that tumor-suppressive effects in mice bearing UCH-L1-depleted tumors were more pronounced compared with those in control mice following administration of paclitaxel, while this suppressive function was abolished by PKM2 overexpression (Fig. 7H-J). Consistently, similar results were observed from IHC staining of Ki67 and TUNEL (Fig. 7K). Therefore, these above data suggest that UCH-L1 knockdown reduces paclitaxel resistance, and this may be attributed to declined PKM2 expression level and restricted glycolysis.Fig. 7. Inhibition of UCH-L1-PKM2 axis-mediated glycolysis reverses paclitaxel resistance.A Dose–response curves of HCC1806 and HCC1806/TaxR cells. The IC50 was determined according to a dose vs. response curve by GraphPad Prism. B Immunoblotting analysis of PKM2 and UCH-L1 in HCC1806 and HCC1806/TaxR cells. C The glucose consumption, lactate and ATP production were determined and compared in HCC1806 and HCC1806/TaxR cells. D HCC1806/TaxR cells from the indicated groups were incubated with PTX for 72 h, and cell viability was measured by CCK-8 assay. E Cell apoptosis of HCC1806/TaxR cells with indicated treatment was measured by TUNEL assay. Scale bars, 150 μm. F HCC1806/TaxR cells with or without UCH-L1 depletion were incubated with 20 nM PTX for 24 h, and cell proliferation was measured by EdU assay. Scale bars, 150 μm. G Colony formation of HCC1806/TaxR cells (300 cells per well) with or without UCH-L1 depletion in the presence of DMSO or 20 nM PTX. HCC1806/TaxR cells were subcutaneously injected into 6-week-old female nude mice and subjected to PTX treatment as planned (n = 5). Tumor sizes were measured and calculated as curves shown (H). The mice tumors of each group were photographed at the termination of the experiment (I). Tumor weights were measured at the end of the experiment (J). Characterization of HCC1806/TaxR xenograft tumors with histologic analysis by IHC staining of Ki67 and TUNEL (K). Data are represented as mean ± SD of biological triplicates. *p < 0.05; **p < 0.01; ***p < 0.001.
UCH-L1 positively correlates with PKM2 in TNBC
To further verify the important role of PKM2 in the paclitaxel resistance of TNBC cells, we evaluated PKM2 expression in TNBC and normal breast tissues by GEPIA2 analysis. As shown in Fig. 8A, the expression of PKM2 was dramatically upregulated in TNBC. Furthermore, a high expression level of PKM2 mRNA was associated with worse overall survival in chemotherapy-treated TNBC (Fig. 8B), indicating that PKM2 is correlate with chemotherapy efficiency. Notably, PKM2 expression was elevated in chemo-resistant patients compared with those who respond well to chemotherapy; and similarly, patients who do not exhibit response to taxane had a significantly higher level of PKM2 expression (Fig. 8C). Moreover, we performed IHC staining of UCH-L1 and PKM2 from 40 cases of TNBC tissues and evaluated the correlation between both (Fig. 8D). The data demonstrated that PKM2 scored higher in high UCH-L1 expression group (Fig. 8E); on the other hand, patients with high PKM2 expression showed high level of UCH-L1 (Fig. 8F). Furthermore, a Pearson’s correlation analysis revealed UCH-L1 and PKM2 positively correlated with each other (Fig. 8G). All of these results indicate the tremendous potential of targeting UCH-L1-PKM2 axis as promising therapeutics for chemotherapy-resistant TNBC.Fig. 8. Positive correlation between UCH-L1 and PKM2 expression in TNBC.A Boxplot of the PKM2 mRNA expression levels in TNBC and normal breast tissues from the GEPIA2 database. B Kaplan-Meier plots for the overall survival of chemotherapy-treated TNBC patients according to PKM2 mRNA expression from the Kaplan–Meier Plotter database. C The PKM2 expressions in non-responder and responder groups in breast cancer cohorts from ROC plotter server. The clinical response was evaluated based on pathological response. D Representative images of IHC of UCH-L1 and PKM2 in 40 TNBC tissues. The correlation analysis between UCH-L1 and PKM2 expression was performed using Fisher’s Exact Test. E Quantitative expression of IHC staining of PKM2 in high or low UCH-L1 groups. F Quantitative expression of IHC staining of UCH-L1 in high or low PKM2 groups. G The positive correlation between UCH-L1 and PKM2. The Pearson correlation test was used. H Mechanistic diagram of UCH-L1-catalyzed PKM2 deubiquitination in the regulation of glycolysis and chemoresistance. Depletion of UCH-L1 led to PKM2 degradation and downregulation, thereby inhibiting glycolysis and re-sensitizing TaxR-TNBC cells to chemotherapy.
Discussion
Resistance to paclitaxel-based chemotherapy is one of the major obstacles contributing to treatment failure of TNBC. Previous studies have reported that enhanced activation of oncogenic molecular signaling pathways like Ras-ERK, PI3K-Akt, IGFR-1, FGFR, EMT and angiogenesis in breast cancer cells promotes tumor progression and chemotherapy resistance [29]. However, because of the high heterogeneity of breast cancer, the molecules and signaling pathways responsible for drug resistance are far of completely understood. In this study, we demonstrated that UCH-L1, a highly active deubiquitinase expressed in TNBC, promoted resistance to chemotherapy, especially paclitaxel resistance. Specifically, patients with high levels of UCH-L1 exhibited worse survival and decreased response to paclitaxel. Moreover, the in vitro anti-tumor effect of paclitaxel was attenuated following UCH-L1 overexpression. More than that, gene silencing of UCH-L1 increased the sensitivity of TNBC to paclitaxel both in vitro and in vivo. Altogether, our results strongly suggest that UCH-L1 may serve as a biomarker for predicting paclitaxel resistance.
Aerobic glycolysis is a well-recognized metabolic characteristic of cancer and it endows cancer cells with advantageous growth abilities and resistance to therapeutic interventions. On the one hand, tumor cells adopted such a metabolic way to produce sufficient energy to support their rapid growth; on the other hand, the intermediate metabolites generated by the high rate of glycolytic flow provided a variety of raw materials for biosynthetic processes [30]. Furthermore, a large amount of lactate secretion formed an immunosuppressive microenvironment conducive to MDSCs, Tregs and M2 TAMs, thus leading to tumor immune escape [31, 32]. In addition, the accumulated lactates in tumor cells promoted resistance to radiotherapy and chemotherapy through lactylation [31, 33]. For the above reasons, targeting the pathway of glycolysis is considered a promising therapeutic strategy. Here, based on our proteome profiles from HCC1806 cells with UCH-L1 knockdown and control cells, we performed GSEA analysis. Interestingly, the analysis of UCH-L1-associated differentially expressed proteins revealed a notable enrichment in Glycolysis/Gluconeogenesis, which prompts us to speculate a possibility that UCH-L1 may promote chemoresistance via manipulation of glycolysis. To verify this hypothesis, we first examined the effect of UCH-L1 on glycolysis. Our observations that UCH-L1 knockdown decreased glucose consumption, lactate and cellular ATP production, while overexpression of UCH-L1 elevated glucose consumption, lactate and ATP production forcefully indicated an important role of UCH-L1 in regulating glycolysis. Moreover, the treatment of 2-DG abolished the paclitaxel resistance effect induced by UCH-L1 overexpression, which confirmed the participation of glycolysis in UCH-L1-enhanced chemoresistance. Of course, there are still some issues deserve further attention. Since UCH-L1 activates glycolysis, whether it functions in remodeling the tumor microenvironment? Another, exploration of the effect of UCH-L1 on global lysine lactylation would help deciphering the multifaced roles of UCH-L1 in cancer.
With the intention of addressing the molecular mechanism underlying UCH-L1-enhanced glycolysis, we conducted IP assay to identify UCH-L1-interacting proteins. Results from LC-MS/MS analysis implied that PKM2 was one of the binding partners of UCH-L1. PKM2 is one of the four isoforms of pyruvate kinase and catalyzes the conversion of phosphoenolpyruvate (PEP) to pyruvate with concomitant production of ATP. Recently, PKM2 has emerged as a key player in a range of cancer biology processes, including metabolic regulation, immunomodulation, inflammation, DNA damage repair and mRNA processing [34]. Therefore, the regulation of PKM2 expression has gained significant attention in cancer research. In this study, we found that UCH-L1 stabilized PKM2 through blocking ubiquitination-mediated degradation of PKM2.
Previous studies have reported that PTMs dictate various structural and functional aspects of PKM2, including its stability, enzymatic activity, oligomeric state, subcellular localization, and interaction with other proteins. For instance, phosphorylation of PKM2 at tyrosine residues such as Y83, Y105, Y148, Y175, Y370 and Y390 affects catalytic activity [35]; serine phosphorylation at S202 and S37 promotes its nuclear translocation; threonine residues such as T454 and T328 phosphorylation hampers the mitochondrial respiration and apoptosis [36]. P300-mediated acetylation of PKM2 at K305 and K433 affects the kinase activity and stability of the protein [13]. In addition, as an important PTM of PKM2, the specific hydroxylation of Pro403 and Pro408 facilitated by prolyl hydroxylase 3 (PHD3) promotes its interaction with the HIF1 transcription complex, and thereby contributing to the transactivation of HIF target genes [37]. Other PTMs of PKM2 like oxidation at Cys358, and glycosylation at Thr405 and Ser406 confer the transition to the dimeric form and promote the Warburg effect [38]. Notably, recent studies have identified the E3 ligase Parkin-induced monoubiquitination of PKM2 at Lys186 and Lys206 residues reduces its enzymatic activity instead of stability [39]. Regarding studies on the regulation of PKM2 stability, CHIP and TRIM58 have been recognized as the E3s of PKM2 that ubiquitinate PKM2 for degradation, leading to inhibition of tumor growth [40, 41]. Moreover, USP4/7/20 have been reported to deubiquitinate and upregulate PKM2 [42, 43]; however, the detailed molecular mechanisms are not particularly clear. Here we report that UCH-L1 is a potential DUB responsible for PKM2 stabilization, as it abolishes PKM2 ubiquitination in a DUB activity-dependent manner. More specifically, the N terminal of UCH-L1 physically interacted with the C terminal domain of PKM2, catalyzed K48-linked deubiquitination of PKM2 at Lys498 residue to prevent proteasome-mediated degradation, thereby resulting in PKM2 stabilization.
The discovery of PKM2 as a substrate for UCH-L1 prompted us to investigate the role of PKM2 in paclitaxel resistance. First, we observed that both UCH-L1 and PKM2 levels were increased in paclitaxel-resistant TNBC cells. Second, cells with UCH-L1 or PKM2 knockdown or treated with glycolysis inhibitor 2-DG showed recovery of sensitivity to paclitaxel. More than that, the re-sensitization effect of UCH-L1 depletion on paclitaxel could be largely reversed by ectopic expression of PKM2. Further analysis based on database data demonstrated that highly expressed PKM2 dictated poor prognosis and limited response in TNBC patients treated with chemotherapy. IHC results of tumor samples from patients with TNBC showed a positive correlation between UCH-L1 and PKM2. Therefore, our data revealed the potential value of targeting UCH-L1-PKM2 axis for overcoming chemoresistance.
In summary, this study identified UCH-L1 as a potential therapeutic target for reversing chemoresistance in TNBC. Mechanistically, highly expressed UCH-L1 in TNBC interacted with the C-terminal domain of PKM2 to deubiquitinate and stabilize PKM2. Up-regulated PKM2 enhanced glycolysis and consequently leading to chemotherapy resistance. Furthermore, inhibition of UCH-L1 improved tumor sensitivity to anti-paclitaxel therapy (Fig. 8H). The findings reported here not only reveal a detailed mechanism by which UCH-L1 promotes PKM2 stabilization and glycolysis, but also provide a rational basis for the UCH-L1/PKM2 axis as a therapeutic target in paclitaxel-resistant TNBC.
Supplementary information
Original Data Supplementary Figures
The reference list from the paper itself. Each links out to its DOI / PubMed record.