DLEU1 promotes cell survival by preventing DYNLL1 degradation in esophageal squamous cell carcinoma

Authenticated ESCC lines TE-1 and KYSE-150 were purchased from the Cell Bank, Type Culture Collection of Chinese Academy of Sciences; KYSE-30 and KYSE-410 were obtained from the Procell Life Science & Technology; EC109 was purchased from the Cell Culture Center of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences. All cells were cultured in RPMI-1640 medium (Gibco) supplemented with 10% fetal bovine serum (BI) at 37 °C with 5% CO₂ in a humidified incubator. All experiments were performed with mycoplasma-free cells.

Human ESCC and paired adjacent noncancerous tissues from 75 ESCC patients who underwent esophagectomy without prior chemotherapy or radiotherapy were obtained from the First Affiliated Hospital, Shihezi University School of Medicine. Subjects were recruited with written informed consent and with approval from the Ethics Committee of the First Affiliated Hospital, Shihezi University School of Medicine.

Total RNA was isolated from cell cultures with a total RNA kit I (Omega Bio-Tek) according to the standard protocol and reverse transcribed into cDNA using SuperQuick RT MasterMix (CWBIO). Target genes were amplified by quantitative real-time PCR (qRT-PCR) with Fast SYBR Green qPCR Master Mix (CWBIO) and specific primers using the 7500 Fast Real-Time PCR System (Applied Biosystems). GAPDH was used as an internal control. The primer sequences used for qRT-PCR are listed in Additional file 1: Table S1.

Cells seeded on 96-well plates (5000/well) were transfected with siRNAs or scramble control RNA oligos. Cell viability was measured using a Cell Counting Kit-8 (CCK-8; Dojindo) following the manufacturer’s protocol. The number of viable cells was quantified according to the absorbance at 450 nm (OD₄₅₀) detected by a microplate reader (BIO‐RAD xMark).

Cells transfected with siRNAs or scramble control RNA oligos were plated on 6-well plates and maintained in medium containing 10% FBS for 10 days for the colony-forming assay. Colonies were fixed with methanol, stained with crystal violet, and counted.

ESCC cells were harvested at 48 h posttransfection by centrifugation with cold PBS and then fixed in ice-cold 70% ethanol at 4 °C overnight. The fixed cells were washed twice with PBS and subjected to PI staining using the Cell Cycle and Apoptosis Analysis Kit (Beyotime Biotechnology). For cell apoptosis analysis, ESCC cells harvested after transfection were stained with the FITC Annexin V Apoptosis Detection Kit I (BD Biosciences) according to the manufacturer's instructions. Acquired data were analyzed using FlowJo software.

RNA pull-down was performed using the Pierce Magnetic RNA–Protein Pull-Down Kit (Thermo Fisher) according to the manufacturer's instructions. Biotin-labeled RNAs were transcribed in vitro with biotin RNA-labeling mix and T7 RNA polymerase (Invitrogen) following the manufacturer’s protocols. The biotinylated sense or antisense DLEU1 RNA generated in vitro was incubated with total cell lysates. Interacting proteins were isolated with streptavidin agarose beads (Invitrogen) at room temperature for 2 h. The complexes were washed briefly with washing buffer three times. The proteins precipitated with streptavidin beads were diluted in protein lysis buffer and resolved by SDS/PAGE, liquid chromatography with tandem mass spectrometry, or immunoblotting.

Equal amounts of total cell lysates were separated by SDS-PAGE and electrophoretically transferred to PVDF membranes. The membrane was blocked with 5% skimmed milk and incubated with primary antibodies against DYNLL1 (ab51603, 1:1000; Abcam), BIM (#12933, 1:1000; Cell Signaling Technology), BCL2 (AB112, 1:1000; Beyotime Biotechnology), β‐actin (1E9A3, 1:1000; ZSGB‐BIO), and FLAG (#14793, 1:1000, Cell Signaling Technology).

Detection of DLEU1 transcripts in formalin-fixed, paraffin-embedded samples was conducted using RNAscope Probe-Hs-DLEU1 and RNAscope 2.5 High Definition Reagent Kit-RED (322350; Advanced Cell Diagnostics) according to the manufacturer's instructions. The staining process was completed by ZSGB-BIO (Beijing, China). Positive staining was indicated by red punctuate dots in the cytoplasm or nucleus. DLEU1 expression levels were categorized into five grades according to the manufacturer’s scoring guidelines: score 0, no staining or < 1 dot per 10 cells; score 1, 1–3 dots per cell (visible at 20–40× magnification); score 2, 4–9 dots per cell; score 3, 10–15 dots per cell and < 10% positive cells have dot clusters; score 4, > 15 dots per cell and > 10% positive cells have dot clusters. Cases that scored higher than 0 were considered positive.

Immunohistochemical staining of DYNLL1 and BCL2 was performed using primary antibodies against DYNLL1 (ab51603, 1:600; Abcam) and BCL2 (AB112, 1:200; Beyotime Biotechnology) on tissue microarray sections by a BOND-MAX Automated IHC/ISH Stainer (Leica). The immunostaining degree of the DYNLL1 and BCL2 proteins was evaluated as previously described [20] by pathologists based on the nuclear staining intensity (intensity score) and percentage of positive cells (extent score). The final immunoreactivity score for each sample was the product of the intensity score and extent score.

Four-week-old athymic female BALB/c mice were housed under specific pathogen-free conditions. For tumorigenicity determination, 1 × 10⁶ cells suspended in a volume of 0.1 ml were subcutaneously injected into the armpits of mice. The mice were sacrificed at 21 days postinjection. For treatment, mice bearing established xenografts of shDLEU1 and shCtrl cells were randomly divided into groups and intraperitoneally injected every other day for two weeks with PBS (vehicle) as a control or cisplatin (1 mg/kg body weight). Tumor volume was measured every other day and calculated using the formula length × width² × π/6. All animal experimental procedures were approved by the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Shihezi University School of Medicine.

Numerical data are presented as the mean ± SEM. Statistical analysis was carried out using GraphPad Prism. Comparisons of changes between two groups were conducted using nonparametric paired or unpaired Wilcoxon tests or parametric unpaired Student's t tests. Comparisons between multiple groups were performed using one-way ANOVA with Tukey’s post-hoc test. Kaplan–Meier analysis and log-rank tests were used for survival estimates. P < 0.05 was considered to indicate statistical significance.

To identify lncRNAs involved in the development of ESCC, we first analyzed two independent lncRNA + mRNA microarray profiles comprising 119 and 60 pairs of ESCC samples with matched adjacent normal tissues, and screened for differentially expressed lncRNAs in ESCC (Fig. 1A). Four of the most frequently overexpressed lncRNAs in cancer tissues, namely, LINC01614, LOC284930, KCNMB2-AS1, and DLEU1, were selected for functional analysis. To evaluate their effect on cell growth, ESCC cells were transfected with a pool of three siRNAs targeting each of these four lncRNAs. Among those, we found that knocking down DLEU1 had a potent suppressive effect on ESCC cell growth (Fig. 1B). Thus, we focused on DLEU1 for further studies. By comparing the two independent cohorts, we found that DLEU1 expression was higher in 118 (99.2%) and 59 (98.3%) tumors than in their matched normal counterparts (Fig. 1C, D). ESCC patients with high DLEU1 expression tended to have decreased overall survival compared with those with low DLEU1 expression (Fig. 1E). We also observed significantly higher DLEU1 expression in tumors than in normal tissues and a correlation between DLEU1 upregulation and worse prognosis in ESCC patients from The Cancer Genome Atlas (TCGA) project (Fig. 1F, G). Moreover, overexpression of DLEU1 was positively correlated with lymph node metastasis and advanced clinical stages (Additional file 1: Table S2). RNAscope assays also confirmed upregulated DLEU1 expression in ESCC compared to adjacent nontumor tissues (Fig. 1H). To gain insight into the underlying mechanism of DLEU1 in ESCC development and progression, we conducted gene set enrichment analysis (GSEA) to identify the potential DLEU1-regulated pathways. Intriguingly, GSEA enrichment plots showed that genes upregulated in breast cancer stem cells, genes upregulated in metastatic prostate cancer, genes highly expressed in liver cancer with poor survival, and genes regulating mitochondrial membrane permeability that are involved in the apoptotic process were significantly enriched in ESCC with high DLEU1 expression levels (Fig. 1I), suggesting the potential contribution of DLEU1 to poor prognosis and mitochondrial apoptosis.

To clarify the biological significance of DLEU1 in ESCC, EC109 and KYSE-410 cells with high endogenous DLEU1 levels (Fig. 6C) were transfected with specific siRNAs to knockdown DLEU1 expression. We confirmed that knockdown of DLEU1 by two different siRNAs resulted in decreased expression levels of DLEU1 in ESCC cells (Additional file 1: Fig. S1). CCK-8 and colony formation assays demonstrated that silencing DLEU1 significantly decreased the growth capacity and colony formation efficiency (Fig. 2A, B). DLEU1 repression also led to a drastic downregulation of the migratory ability and invasive potential of ESCC cells (Fig. 2C). In contrast, ectopic expression of DLEU1 in KYSE-150 cells with low endogenous DLEU1 levels resulted in a significant enhancement of cell growth (Fig. 2D), migration, and invasion (Fig. 2E). When transplanted into nude mice, the tumor growth of cells with DLEU1 overexpression was significantly promoted compared with that of control cells (Fig. 2F–H). Collectively, these data suggest a growth-promoting role of DLEU1 in ESCC.

The enhancement of growth induced by DLEU1 could reflect either an increased proliferation rate or protection from cell apoptosis. To determine how DLEU1 contributes to the growth of ESCC cells, we first examined the effect of DLEU1 knockdown on cell proliferation using PI staining. No significant changes in cell cycle distribution were observed after DLEU1 knockdown (data not shown), suggesting that silencing DLEU1 had little impact on cell cycle progression. Nevertheless, the annexin V/PI assay showed that knockdown of DLEU1 by shRNAs resulted in a marked increase in cell apoptosis rates in KYSE-410 cells, whereas DLEU1 overexpression inhibited apoptotic cell death in KYSE-150 cells (Fig. 2I, J). The antiapoptotic effect of DLEU1 was further confirmed by mitochondrial membrane potential detection and TUNEL staining (Fig. 2K, L). The results indicated that DLEU1 was required for the survival and tumorigenicity of ESCC cells and played this role by inhibiting apoptosis.

A number of lncRNAs have been reported to exert their biological function by forming complexes with proteins [21]. We therefore performed an RNA pull-down assay followed by mass spectrum analysis to identify DLEU1-associated proteins and elucidate the potential mechanisms of DLEU1 in ESCC. Among the identified DLEU1-binding proteins, we selected DYNLL1 based on the functional annotation of precipitated proteins (Fig. 3A). Dynein light chain LC8-type 1 (DYNLL1, also known as DLC1 and LC8) is a hub protein that has been proposed to regulate mitochondrial apoptosis [22]. The association of DYNLL1 with DLEU1 sense RNA, but not antisense RNA, was further verified by immunoblotting (Fig. 3B). We also performed RNA immunoprecipitation analysis of total cell lysates with the anti-DYNLL1 antibody. qRT-PCR analysis of the immunoprecipitates demonstrated that DYNLL1 was associated with DLEU1 RNA (Fig. 3C). To identify which region was required for the association of DLEU1 with DYNLL1, we constructed a series of DLEU1 deletion mutants named Del1 (1–1086 nt), Del2 (1–724 nt), Del3 (725–1448 nt), and Del4 (363–1448 nt). RNA fragments were transcribed in vitro from these deletion mutants and used for RNA pull-down assays. Immunoblotting analysis of the DYNLL1 protein showed that the DLEU1 construct Del3 nearly completely lost its ability to bind DYNLL1, suggesting that the region was not essential for DLEU1 binding to DYNLL1 (Fig. 3D).

Next, we explored whether DLEU1 could regulate the expression of DYNLL1 in ESCC cells. No significant change in DYNLL1 transcription was observed in DLEU1-overexpressing KYSE-150 cells compared with vector control cells (Fig. 3E). Nevertheless, DYNLL1 protein levels were markedly increased after ectopic DLEU1 expression in KYSE-150 cells but drastically decreased following knockdown of DLEU1 by siRNAs or shRNAs in EC109 and KYSE-410 cells (Fig. 3F, G), suggesting that DLEU1 could enhance DYNLL1 expression by posttranscriptional mechanisms. The protein synthesis inhibitor cycloheximide (CHX) was used to evaluate the effect of DLEU1 on the degradation of DYNLL1. We found that knockdown of DLEU1 shortened the half-life of DYNLL1 in EC109 and KYSE-410 cells (Fig. 3H). Moreover, inhibition of proteasome activity with MG132 prevented the downregulation of DYNLL1 protein caused by DLEU1 silencing (Fig. 3I). We found that ectopic expression of DLEU1 in KYSE-150 cells increased the stability of DYNLL1 by suppressing its ubiquitination (Fig. 3J). In contrast, a higher ubiquitination level of DYNLL1 was observed in cells expressing DLEU1-Del3, which was unable to associate with DYNLL1 (Fig. 3D), compared with the full-length and other deletion fragments of DLEU1 (Fig. 3K). Together, these results demonstrated that DLEU1 increased the stability of DYNLL1 by inhibiting its ubiquitin–proteasome-mediated degradation.

Ubiquitin ligase E3, the critical factor of the ubiquitin–proteasome system for selective degradation, can recognize proteins that will be degraded and attach ubiquitin to target proteins [21, 23]. To identify the ubiquitin ligase targeting DYNLL1 in ESCC cells, we reanalyzed the DLEU1 mass spectrometry data and searched for coprecipitated E3 ubiquitin ligases. We selected RNF114 as the candidate since its coverage in the DLEU1 group was markedly lower than that in the DLEU1 antisense group (Fig. 3A). Pull-down assays revealed that RNF114 coprecipitated with DYNLL1 and vice versa (Fig. 4A, B), suggesting that RNF114 is the ubiquitin ligase targeting DYNLL1 for degradation. Indeed, overexpression of RNF114 led to an increased ubiquitination level of DYNLL1 (Fig. 4C). Furthermore, DLEU1 overexpression abolished the ubiquitination of DYNLL1 induced by RNF114 and restored DYNLL1 expression (Fig. 4D). We found that DLEU1 knockdown abrogated siRNF114-induced DYNLL1 expression (Fig. 4E). Conversely, ectopically expressed DLEU1 restored the RNF114-induced degradation of DYNLL1 (Fig. 4F). These results indicated that RNF114 mediated DYNLL1 ubiquitination and degradation, while DLEU1 could interfere with the degradation of DYNLL1 mediated by RNF114.

To identify specific sites where the DYNLL1 protein is ubiquitinated, we generated mutant derivatives of DYNLL1 in which Lys-9, -31, -36, -43, or -49 were replaced by Arg (DYNLL1-K9R, -K31R, -K36R, -K43R, or K49R; Fig. 4G) and examined their ubiquitination in the presence of ectopic RNF114 in ESCC cells. All DYNLL1 derivatives underwent RNF114-mediated ubiquitination except for DYNLL1-K36R and DYNLL1-K43R, whose ubiquitination was attenuated (Fig. 4H). In addition, we found that overexpression of RNF114 did not induce the degradation of DYNLL1-K36R or DYNLL1-K43R (Fig. 4I). These results indicated that DYNLL1 might be ubiquitinated at Lys-36 and Lys-43 by RNF114.

DYNLL1 has been proposed to limit the proapoptotic activity of BIM and contribute to the efficient expression of the antiapoptotic protein BCL2 [22]. We therefore asked whether this cascade is involved in the modulation of apoptotic cell death by DLEU1 in ESCC. Immunoblotting demonstrated that the level of BCL2 was strongly downregulated, while BIM expression was somewhat increased in KYSE-150-DLEU1 cells upon DYNLL1 knockdown (Fig. 5A, B). In contrast, ectopic DLEU1 expression upregulated the expression of the antiapoptotic protein BCL2 (Fig. 5C). The residues Cys2 and Ser88 of DYNLL1 have been implicated in affecting its interaction network with other factors [24, 25]. To illustrate the functional sites of DYNLL1, mutation of Cys2 to alanine (C2A), phospho-null mutation of Ser88 (S88A), and phosphomimetic Ser88 mutation (S88D) constructs were generated. Immunoblotting analysis revealed that DYNLL1-S88D resulted in slightly higher BIM expression but decreased BCL2 expression compared with wild-type DYNLL1 and other mutants in KYSE-150 cells (Fig. 5D). siDYNLL1-induced changes in BCL2 and BIM expression levels could be reversed by BIM knockdown (Fig. 5E). Consistently, DYNLL1 knockdown mitigated the promoting effect of DLEU1 on cell survival, while the concomitant loss of BIM completely abolished the effect of DYNLL1 silencing on apoptotic cell death (Fig. 5F).

To establish whether the expression levels of DLEU1, DYNLL1, and BCL2 are clinically associated, we evaluated DLEU1 RNA expression by RNAscope and evaluated the protein levels of DYNLL1 and BCL2 by immunohistochemistry staining in ESCC tissues. DYNLL1 and BCL2 protein levels were positively correlated with DLEU1 levels in ESCC samples (r = 0.567 and 0.370). A positive correlation between DYNLL1 and BCL2 levels was also observed (r = 0.313; Fig. 5G). Consistently, histological examination of xenograft tumors in nude mice demonstrated higher DYNLL1 and BCL2 protein levels in the DLEU1-overexpressing group than in the control group (Fig. 5H). Collectively, these data demonstrated that DLEU1/DYNLL1 axis-mediated upregulation of BCL2 was involved in the survival of ESCC cells.

To decipher the underlying mechanisms of dysregulated DLEU1 expression in ESCC, we explored tumor-intrinsic oncogenic events in the DLEU1 gene locus. Genetic alterations of the DLEU1 locus in the TCGA-ESCC cohort were first analyzed (Additional file 1: Fig. S2A). The DLEU1 gene showed rare alterations (2.1%, 2/96 cases with deep deletions). A positive correlation of DLEU1 RNA expression with its gene copy number was observed in ESCC samples, suggesting that the upregulation of DLEU1 was partly due to DNA copy number gain (Additional file 1: Fig. S2B, C). DNA methylation was found to be one of the causes of abnormal gene expression, which promoted the genesis of tumors. Intriguingly, DLEU1 expression was negatively correlated with the DNA methylation level of global or individual CpG sites within the DLEU1 promoter region in ESCC patients (Fig. 6A). In addition, Kaplan–Meier analysis revealed that ESCC patients with low DLEU1 methylation had significantly shortened overall survival and disease-specific survival compared with those with high DLEU1 methylation (Fig. 6B). Treatment with the DNA methyltransferase inhibitor decitabine significantly enhanced DLEU1 expression in a series of ESCC cell lines (Fig. 6C). Consistently, DNA demethylation by decitabine enhanced DYNLL1 and BCL2 protein levels in EC109 and KYSE-410 cells (Fig. 6D). These findings collectively indicated that the upregulation of DLEU1 in ESCC was facilitated by DNA copy number gain and promoter hypomethylation.

In line with our above finding that DLEU1 could prevent apoptotic cell death, knockdown of DLEU1 enhanced the antitumor effect of cisplatin treatment in KYSE-410 cells (Fig. 7A). To further explore the biological significance of DLEU1 in supporting tumor development, mice with established xenografts with or without DLEU1 silencing were randomized to receive treatment with vehicle control or cisplatin (Fig. 7B). DLEU1 knockdown or cisplatin treatment alone was sufficient to impair tumor growth. Moreover, compared with cisplatin alone, the combination of cisplatin and DLEU1 knockdown exhibited a stronger inhibitory effect on xenograft tumor growth in nude mice (Fig. 7C, D). Histological examination of xenograft tumors showed that DLEU1 silencing was accompanied by decreased DYNLL1 and BCL2 protein levels (Fig. 7E). These results suggested that targeting DLEU1 could increase the sensitivity of ESCC cells to cisplatin therapy.