USP29 enhances chemotherapy-induced stemness in non-small cell lung cancer via stabilizing Snail1 in response to oxidative stress

Abstract

Chemotherapy remains an essential part of diverse treatment regimens against human malignancies. However, recent progressions have revealed a paradoxical role of chemotherapies to induce the cancer stem cell-like features that facilitate chemoresistance and tumor dissemination, with the underlying mechanisms underinvestigated. The zinc-finger transcription factor Snail1 is a central regulator during the epithelial-mesenchymal transition process and is closely implicated in cancer progression. Snail1 expression is strictly regulated at multiple layers, with its stability governed by post-translational ubiquitylation that is counterbalanced by the activities of diverse E3 ligases and deubiquitylases. Here we identify the deubiquitylase USP29 as a novel stabilizer of Snail1, which potently restricts its ubiquitylation in a catalytic activity-dependent manner. Bioinformatic analysis reveals a reverse correlation between USP29 expression and prognosis in lung adenocarcinoma patients. USP29 is unique among Snail1 deubiquitylases through exhibiting chemotherapy-induced upregulation. Mechanistically, oxidative stresses incurred by chemotherapy stimulate transcriptional activation of USP29. USP29 upregulation enhances the cancer stem cell-like characteristics in lung adenocarcinoma cells to promote tumorigenesis in athymic nude mice. Our findings uncover a novel mechanism by which chemotherapy induces cancer stemness and suggest USP29 as a potential therapeutic target to impede the development of chemoresistance and metastasis in lung adenocarcinoma.

Fig. 1. USP29 is a novel Snail1 regulator that potently stabilizes its levels.

a HEK293T cells were co-transfected with constructs expressing Flag-Snail1 and various GFP-DUBs as indicated. After 24 h, the expression levels of Snail1 and DUBs were assessed by Western blotting. Tubulin was probed to confirm equal loading. Images are representative from three independent experiments. b Densitometry analysis of Snail1 band intensities from A summarizing three biological repeats, with pEGFP-C1 control set as 1. The data points were analyzed by Gaussian distribution showing mean (solid line) and 95% Confidence Interval (dashed lines). c HEK293T cells were co-transfected with Flag-Snail1 expressing plasmid at constant amounts and pEGFP-USP29 construct with escalating doses ranging from 0 to 1 μg, prior to Western blotting analysis with indicated antibodies. Below column chart shows the relative quantification of Snail1 expression (n = 3). d HEK293T cells co-transfected with Flag-Snail1 and pEGFP-C1 control or pEGFP-USP29 constructs were treated with the proteasomal inhibitor bortezomib (200 nM) for 3 and 6 h before cell lysis and Western blotting analysis using indicated antibodies. Below column chart shows the relative quantification of Snail1 levels. All error bars represent the standard error of the mean (SEM), with *p-value < 0.05 and n.s. representing not significant.

Fig. 2. USP29 interacts with Snail1 and requires catalytic USP domain to stabilize Snail1.

a HEK293T cells were co-transfected with Flag-Snail1 vector along with Flag-HA-GFP, Flag-HA-USP28, or Flag-HA-USP29 constructs for 24 h. Cells were treated with cycloheximide (100 μg/ml) to stop protein synthesis for indicated times and lysed for Western blotting assays with indicated antibodies. b turnover curves show the relative amounts of Snail1 at each time point from (a) (n = 3). c schematic illustration of the domain structure in USP29 and corresponding truncations with indicated amino acid sequences. d Flag-Snail1 and various GFP-USP29 constructs were co-transfected into HEK293T cells for 24 h, and then cell lysates were analyzed with Western blotting using indicated antibodies. Below column chart shows the relative quantification of Snail1 levels (n = 3). e HEK293T cells were co-transfected with indicated combinations of Flag-Snail1 and diverse USP29 constructs for 24 h. GFP or GFP-tagged USP29 variants were immunoprecipitated with anti-GFP antibodies. Samples were processed for Western blotting assays to probe for co-immunoprecipitated Snail1. f Flag-tagged Snail1 was immunoprecipitated from HEK293T cell lysates following co-transfections of Flag-Snail1 with pEGFP, pEGFP-USP28, or pEGFP-USP29-USP constructs. Immunoprecipitates were analyzed by Western blotting using indicated antibodies. All error bars represent the standard error of the mean (SEM), with *p value < 0.05 and n.s. representing not significant.

Fig. 3. The Snail1-stabilizing effect of USP29 depends on its catalytic USP domain-mediated deubiquitylation.

a U2OS cells were co-transfected with Flag-Snail1 and various GFP-USP29 constructs before immunofluorescence analysis. Flag-Snail1 was stained using anti-Flag antibody. GFP-USP29 variants were visualized through GFP. Micrographs show representative images from each condition captured under the same intensity settings, with DAPI staining showing nuclei. Scale bar = 10 µm. b, c HEK293T cells were co-transfected with Flag-Snail1 and indicated control or GFP-USP29 plasmids. Following bortezomib treatment (500 nM, 6 h) to block proteasomal degradation, Flag-Snail1 was immunoprecipitated from cell lysates using anti-Flag antibody. Immunoprecipitates were analyzed by Western blotting to examine the ubiquitylation status of Flag-Snail1 in different groups. Column charts show the relative quantification of the ubiquitin signal. d HEK293T cells were co-transfected with Flag-Snail1 (0.5 μg) and GFP-USP29-C294A (0–1 μg) constructs for 24 h. Cell lysates were analyzed by Western blotting with indicated antibodies. Below column chart shows the quantification data of Snail1 expression (n = 3). All error bars represent the standard error of the mean (SEM), with *p value < 0.05.

Fig. 4. USP29 expression correlates with enhanced cancer stem cell features and poor prognosis in lung cancer.

The prognosis data of lung cancer patients were retrieved from the KM-plotter website (www.kmplot.com), showing correlations between USP29 expression and overall survival in total lung cancer (a), lung squamous carcinoma (LUSC, b), and lung adenocarcinoma (LUAD, c). Patients were split by medians of USP29 expression. d RNA-seq data of indicated DUBs in 517 LUAD patients from TCGA (The Cancer Genome Atlas) were retrieved from the cBio Cancer Genomics Portal (http://cbioportal.org). Mean RPKM (reads per kilobase of exon per million reads mapped) of individual DUB genes was calculated to enable the quantitative comparison. e Characterization of H1299 and H1975 cells stably expressing USP29 in parallel with vector control stable cell lines. f Representative images from transwell assays using H1299 and H1975 cell lines stably transfected with pCDH and pCDH-USP29 as indicated. g Column charts show quantification data of migrated cells per microscopic view on average from three biologically independent experiments. (h), representative images of spheroids formed by H1299 and H1975 stable cell lines after day 8 and 15. Scale bar = 50 μm. i Column charts show relative quantification of spheroid numbers compared to pCDH control groups for H1299 and H1975 cells stably expressing USP29, respectively (n = 3). All error bars represent the standard error of the mean (SEM), with *p value < 0.05 and **p value < 0.01.

Fig. 5. USP29 expression enhances stemness of lung adenocarcinoma cells and in vivo tumorigenesis in xenograft mouse model.

a, b Side population was identified in H1299 and H1975 cells stably transfected with control or USP29-expressing vectors with the presence and absence of verapamil (200 μM). Scatter plots show representative flow cytometric analysis from three independent experiments for H1299 and H1975 as indicated, with solid boxes indicating side population cells. c, d The relative quantification of side population cells from (a, b), respectively (n = 3). e The capabilities of H1299-pCDH and H1299-pCDH-USP29 stable cell lines to form tumor xenograft in athymic nude mice were compared in parallel, with the image showing resected xenograft tumors. f The volumes of tumor xenograft formed by two cell lines during the experimental period were measured and plotted. g Weights of tumor xenografts resected were measured and plotted. h Tumor xenograft tissues were processed for Western blotting analysis to probe for USP29 and Snail1 levels. Tubulin blot shows equal loading. i Quantification data from (h). All error bars represent the standard error of the mean (SEM), with *p value < 0.05 and **p value < 0.01.

Fig. 6. USP29 is transcriptionally activated in response to chemotherapy and oxidative stress in lung adenocarcinoma cells.

a–d A549 and H1299 cells were treated with doxorubicin (DOX, 5 μg/ml) and taxol (10 μM) separately as indicated for 3 and 6 h, prior to RNA extraction and cDNA preparation by RT-PCR. Relative mRNA levels of USP29, DUB3, OTUB1, and USP27X were assessed by carrying out PCR assays with specific primer sets. GAPDH mRNA was detected as control. Images on the left show representative results from three independent experiments. Column charts on the right show quantification data of the relative mRNA amounts of the four DUBs (n = 3). e, f Cultured A549 and H1299 cells were treated with hydrogen peroxide at 2 mM for 1 h before recovery in normal culture media. After indicated times, cells were harvested to extract total RNAs. cDNA was prepared by reverse transcription and used in PCR assays to examine the mRNA levels of USP29 and Snail1 at various conditions. Images show representative results from three independent assays, with column charts showing the quantification data of relative USP29 mRNA abundance (n = 3). All error bars represent the standard error of the mean (SEM), with *p value < 0.05 and **p value < 0.01.

Fig. 7. Schematic diagram depicting the oxidative stress-USP29-Snail1 regulation axis of cancer cells in responsive to chemotherapies.

Our current working model proposes that lung adenocarcinoma cells display their intrinsic capabilities of rewiring transcriptional circuitry to increase USP29 expression in responsive to oxidative stresses incurred by chemotherapeutic agents. USP29 upregulation leads to the stabilization of Snail1, which acts as the master regulator to enable the expression of a range of mesenchymal genes. As a consequence, this oxidative stress-USP29-Snail1 regulation axis confers cancer cells enhanced stemness and chemoresistance. TF transcription factor, Ub ubiquitin, E3 ubiquitin E3 ligase.