Exosome-delivered pepDCBLD2 promotes glucose metabolism to induce chemoresistance in colon cancer Authors: Xinyi Wang, Jiayu Yang, Haiou Yang, Yue Xu, Huiya Wang, Mengyu Sun, Ming Gao, Ye Wang, Haiyang Zhang Published: 2026 ABSTRACT AbstractBackgroundOxaliplatin is a standard treatment for colon cancer, but chemoresistance is a considerable challenge. Dysregulation of circRNAs is involved in cancer and translatomics has found that some circRNAs encode polypeptides.MethodsRibosome footprint profiling combined with circRNA sequencing was applied to screen circRNA with coding potential. Immunoprecipitation and mass spectrum were performed to explore the potential mechanism for regulating the downstream pathway. Ultracentrifugation and commercial kits were utilized to extract exosomes. Different kits were used to detect metabolic indexes. CCK-8, immunofluorescence and flow cytometry were conducted for the examination of oxaliplatin sensitivity. A xenograft model in NCG mice was established to demonstrate the pathway in vivo.ResultscircDCBLD2 could encode pepDCBLD2 through rolling circle translation. pepDCBLD2, instead of the circRNA itself, participated in competitive binding to ubiquitin protein ligases, and upregulated the expression of the parent gene DCBLD2. Additionally, exosomes from drug resistant cells transferred pepDCBLD2 to sensitive cells, where it intercellularly regulated the DCBLD2/ERK/PKM2 pathway. This enhanced the glucose metabolism and reactive oxygen species scavenging capacity, ultimately promoting chemoresistance.ConclusionsExosome-delivered pepDCBLD2 encoded by circDCBLD2 promoted glucose metabolism to induce oxaliplatin resistance through the DCBLD2/ERK/PKM2 axis, which might be a potential therapeutic target for future treatment of chemoresistant colon cancer.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12967-026-08876-7.Keywords: circRNA, Polypeptide, Exosome, Glucose metabolism, Chemoresistance Introduction Colorectal cancer ranks third in both incidence and mortality among the most common malignant tumors worldwide [1]. Currently, the treatment options for colorectal cancer include surgery, chemotherapy, targeted therapy and immunotherapy. However, for patients with advanced cancer, treatments are limited and the survival ratio is low, which is usually the result of chemotherapy resistance [2]. Therefore, further exploration of the mechanism of chemoresistance in colorectal cancer may provide novel directions for the clinical treatment of drug-resistant colorectal cancer. Oxaliplatin is a classic chemotherapy drug used for colon cancer. Oxaliplatin has dose-dependent toxicity, which may interrupt treatment, especially in cancers with oxaliplatin resistance. Therefore, it is worth exploring potential methods to sensitize oxaliplatin and reduce the dosage. Studies have shown that oxaliplatin can indirectly cause DNA oxidative damage by inducing the production of reactive oxygen species (ROS) [3–5]. The reduced glutathione (GSH) is an important antioxidant that participates in the clearance of ROS in cells. The maintenance of its reduced state is related to the product NADPH of the pentose phosphate pathway [6, 7]. Many solid tumors rely on rapid aerobic glycolysis [8] for energy supply. Simultaneously, they accumulate a large number of intermediate products to participate in the pentose phosphate pathway, generating more NADPH and maintaining the redox homeostasis within the cells. Due to the heterogeneity of cancer [9], there are often cells with sensitivity and tolerance to chemotherapy, as well as cells with imbalanced metabolism and homeostasis within the tumor. Studies have shown that tumor cells can interact with other tumor cells and various stromal cells in the microenvironment [10]. Exosomes have been proven to act as carriers to mediate signal transduction between cells [11]. Exosomes are single-layer membrane vesicles secreted by cells, with a diameter of approximately 30–100 nm, carrying components such as microproteins, polypeptides, and various non-coding RNAs (ncRNA) [12]. Exosomes, through the bloodstream, act on nearby or distant tissues, regulating various processes such as drug resistance [13], immune evasion [14], invasion and metastasis [15]. However, there are few reports on the study of exosomes carrying polypeptides to regulate chemoresistance. In recent years, numerous studies have reported the structures and functions of ncRNAs. Among them, circular RNAs (circRNAs) can interact with other RNAs or proteins, and participate in regulating signaling pathways [16]. Their expression disorder may lead to various diseases, including malignant tumors [17]. However, only a portion of circRNAs have completely matching RNA sequences and spatial structures. Therefore, other regulatory mechanisms that do not directly rely on base sequences are worthy of exploration [18]. With the development of sequencing technology, proteomic and translatomic studies have found that some so-called ncRNAs contain small open reading frames (ORFs) that can encode polypeptide [19]. These polypeptides regulated parent gene-related proteins or other target proteins, thereby affecting viral infection [20], cancer progression and therapy [21]. However, it is not clear whether circRNAs can regulate oxaliplatin resistance in colon cancer by encoding polypeptides. This study aimed to explore the mechanism by which exosomes from chemoresistant cells carried circRNA-encoding polypeptides to regulate oxaliplatin resistance intercellularly in colon cancer. Through translatomic and transcriptomic sequencing, circDCBLD2 capable of encoding oxaliplatin resistance-related polypeptides was screened out. It was verified that it could undergo rolling circle translation to produce the polypeptide pepDCBLD2 and upregulate the protein level of the parent gene DCBLD2. Subsequently, by protein profiling analysis, it was discovered that DCBLD2 could activate the ERK/PKM2 pathway, promote the nuclear translocation of PKM2, and upregulate the expression level of PKM2. Further research confirmed that the resistant-cell exosomes highly expressed pepDCBLD2 and could deliver pepDCBLD2 to sensitive cells. The transferred pepDCBLD2 increased the expression of DCBLD2, activated the ERK/PKM2 pathway, promoted glucose metabolism and led to oxaliplatin resistance within colon cancer cells. The results of this study revealed a novel mechanism by which exosomes from chemoresistant cells transported circRNA-encoding polypeptides, regulating the DCBLD2/ERK/PKM2 pathway in sensitive cells to promote drug resistance in colon cancer. This provides potential therapeutic targets for improving the treatment of chemoresistant colon cancer in the future. Methods Animals Female NCG mice (BALB/c-nu, 4 weeks) purchased from GemPharmatech Co., Ltd. (Jiangsu, China) were fed in a specific pathogen-free animal facility and allowed to eat and drink ad libitum. Human samples All human samples were obtained from Tianjin First Central Hospital. All patients were confirmed histologically as colon adenocarcinoma and staged as stage III-IV according to the AJCC Cancer Staging Manual (8th edition), receiving first-line chemotherapy primarily based on oxaliplatin. Efficacy was assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST version 1.1). Patients with partial response (PR) or progressive disease (PD) had venous blood collected fasting in the early morning, followed by immediate centrifugation to separate serum. Some tissue sections of the above drug resistant or sensitive patients were also obtained through surgery. Baseline characteristics of the two groups were balanced and comparable (Table 1 in Supplementary Material 6). Cell culture and transfection The human colon cancer cell line SW480 and HCT116 and the human embryo kidney epithelial cell line HEK293T were acquired from the cell bank of the Chinese Academy of Sciences (Shanghai, China). HEK293T cells were cultured in DMEM, and SW480 and HCT116 cells were cultured in 1640 medium. All the cell lines were recently authenticated by STR profiling and tested for mycoplasma contamination. Short interfering RNAs (siRNA) targeting the back-splice junction of circDCBLD2 (RiboBio, China; target sequence: 5’-GATAAACAAGGTGATGGAT-3’) was applied to knock down circDCBLD2. Other siRNAs were also purchased from Guangzhou RiboBio Co., Ltd. All the overexpression (OE) plasmids were purchased from Shanghai GeneChem Co., Ltd. (China). Protein extraction and western blotting (WB) Proteins were extracted with SDS lysis buffer and detected by WB using anti-DCBLD2 (R&D Systems, AF6269; or proteintech, 13168-1-AP; or ImmunoWAY, YN3401), anti-Flag (Abcam, ab205606), anti-MIB1 (ImmunoWAY, YT5815 or YM960008), anti-ERK1/2 (ImmunoWAY, YM8336), anti-phosphoERK1/2 (ImmunoWAY, YP1197), anti-PKM2 (ImmunoWAY, YM8043), anti-phosphoPKM2 (ImmunoWAY, YP1444), anti-CD63 (Santa Cruz, sc-5275), anti-TSG101 (Santa Cruz, sc-7964) and anti-calnexin (Abcam, ab213243). An anti-β-actin antibody (Santa Cruz, sc-47778) was utilized for normalization. WB data were quantified using Image J2x. Co-immunoprecipitation (Co-IP) and LC-MS/MS Co-IP was conducted using anti-Flag (Santa Cruz, sc-166355) and anti-MIB1 (Santa Cruz, sc-393811). The antibody-to-lysate ratio was 2 µg per 500 µg of total protein. The collected protein complexes were subjected to SDS-PAGE, and the gels were analyzed via LC-MS/MS by Shanghai APTBIO Biotechnology Co., Ltd. Exosome isolation and validation Exosomes from cultured cells were extracted by differential ultracentrifugation (detailed in Supplementary Material 6). Serum exosomes were isolated by applying the Total Exosome Isolation Kit (Invitrogen). Exosomes were photographed with an FEI Tecnai T20 transmission electron microscope (Thermo). Sizes and numbers of exosomes were tracked via the NanoSight NS300 system (NanoSight Technology), and all the data were analyzed by the NTA analytical software (Thery and Witwer). PKH26 Red Fluorescent Cell Linker Kit (Sigma) was utilized for exosome staining, and the labeled exosomes were then coincubated with recipient cells for 2–24 h before imaging. Cell apoptosis assay Cells were preprocessed and incubated with 30 µg/mL of oxaliplatin for 48 h. This drug concentration was chosen according to the CCK-8 results and had a relatively small impact on the basic growth of cells. And at 30 µg/mL of oxaliplatin, fluctuations of the drug or some intervention measures might significantly alter the cell activity. Subsequently, both the attached and floating cells were collected and labeled with Annexin V-APC and 7AAD (Lianke Bio). Apoptotic cells were quantified by a fluorescence-activated cell sorting flow cytometer (BD Biosciences). Immunofluorescence (IF) Cells were cultured in 12-well plates with glass coverslips, incubated with 30 µg/mL of oxaliplatin for 24 h and subjected to IF using anti-Flag (Abcam, ab205606), anti-MIB1 (ImmunoWAY, YT5815), phospho-H2AFX-S139 antibody (ABclonal, AP0687) and anti-PKM2 (ImmunoWAY, YM8043). Cellular nuclei were stained with DAPI (Solarbio). Detection of metabolic indicators The relative glucose uptake was assessed by measuring the glucose concentration in medium (Johnson & Johnson), and the lactate production in the medium was detected using a Lactate Assay Kit (Solarbio). A CellTiter-Glo Luminescent Cell Viability Assay (Promega) was applied for ATP measurement. Intracellular NADP + and NADPH were isolated with an acid or alkaline extraction reagent and tested by a CoenzymeⅡNADP(H) Content Detection Kit (Solarbio). GSH contents were measured using a GSH Assay Kit (Solarbio). All the metabolic indicators were calibrated with the number of cells or fresh weight of tumors. For ROS detection, cells were incubated with DCFH-DA (Solarbio) for 30 min, and analyzed by a fluorescence-activated cell sorting flow cytomete. Establishment of tumors in NCG mice A total of 1 × 107 HCT116 cells were injected into the hypodermic areas in the left inguinal regions of the mice in every group randomly. Statistical analysis All the data in our study were based on at least three independent experiments and presented as the mean ± SEM. P < 0.05 was regarded as statistically significant via Chi-square test, t-test or Dunnett t-test: *P < 0.05 and **P < 0.01. Other methods Real-time quantitative polymerase chain reaction (RT-qPCR), Exosome isolation by differential ultracentrifugation, IRES score, Mutant sequence of circRNA plasmids, and Oxaliplatin concentration setting were described in the Supplementary Material 6. Results High-throughput sequencing screening of potential translated circRNAs correlated with chemoresistance To explore whether there are circRNAs that can regulate oxaliplatin resistance in colon cancer by encoding polypeptides, this study simultaneously performed transcriptome and translationome sequencing using oxaliplatin-resistant colon cancer cells HCT116/L-OHP and the corresponding sensitive cells HCT116 (Fig. 1A). Transcriptome sequencing identified a total of 10,961 circRNAs (Fig. 1B), among which 2,085 circRNAs showed significant transcriptional differences (Fig. 1C). The criterion used to define the difference was that P < 0.05, and the fold change was more than twice. Translationome ribosome profiling sequencing output circRNA-specific ORFs (Fig. 1D), namely cORFs spanning the junction reads. In this study, only 121 circRNAs were detected to have reads at the junction sites (Fig. 1E, Supplementary Material 4 ), suggesting that these 121 circRNAs might encode polypeptides through cORFs spanning the junction sites and thereby exert regulatory functions. All circRNA sequences were scored for IRES using IRESfinder in a window-based manner (detailed in Supplementary Material 6), and circDCBLD2 had an IRES score of 0.902581, indicating high reliability (Fig. 1F). These circRNAs were compared with the transcriptome data and a nine-quadrant plot was drawn (Fig. 1G). hsa_circ_0066631 showed no significant difference in the transcriptome, but was significantly upregulated in the translationome with relatively high expression abundance (indicated in Fig. 1E and G). So it was selected for further study. Since hsa_circ_0066631 is derived from the reverse splicing of exons of its parent gene DCBLD2, it was named circDCBLD2 in this study (Fig. 1H). According to the translationome sequencing results, the polypeptide pepDCBLD2 encoded by circDCBLD2 has a molecular weight of 49.918 kDa and a total length of 450 amino acids (aa). Actually, this molecular weight corresponds to a protein size. However, this study only focused on the primary structure, namely the amino acid sequence, without exploring its secondary or tertiary structures. Therefore, it was referred to as a polypeptide in this study. The circDCBLD2 undergoes rolling circle translation spanning the junction site, and the pepDCBLD2 is the repetitive sequence of 69–190 aa in the DCBLD2 protein (Fig. 1I). So the antibody targeting 67–528 aa region of DCBLD2 (including 69–190 aa) was used to recognize the possible endogenous pepDCBLD2 by WB based on size discrimination. In conclusion, circDCBLD2 was screened out with potential coding ability related to oxaliplatin resistance through high-throughput sequencing. High-throughput sequencing of potential translated circRNAs correlated with chemoresistance. A. Schematic diagram of screening of potential coding circRNAs. B. Scatter diagram of circRNA-seq in HCT116/L-OHP and HCT116 cells. The criterion used to define the difference was that P < 0.05, and the fold change was more than twice. C. Bar chart of differential expression of circRNAs. D. Schematic diagram of ribosome footprint sequencing. E. Scatter diagram of circRNA Ribo-seq in HCT116/L-OHP and HCT116 cells. F. Bar chart of IRES score distribution. G. Nine quadrant diagram of translatome and transcriptome correlation analysis. The x- and y-axes represent the fold-chance ratio of the chemo-resistant compared to -sensitive cells. H. Schematic diagram of circDCBLD2 and DCBLD2 mRNA. I. Schematic diagram of pepDCBLD2 and DCBLD2 protein. J. Schematic diagram of two circDCBLD2 OE plasmids of circDCBLD2-Flag fusion and the corresponding insertion mutant. K. Compared to the control, circDCBLD2 was upregulated with either circDCBLD2-Flag OE plasmid transfection by qPCR. L. WB confirmed pepDCBLD2-Flag was encoded in HEK293T cells only transfected with circDCBLD2-Flag OE plasmid. M. WB confirmed pepDCBLD2-Flag was encoded in HCT116 and SW480 cells only transfected with circDCBLD2-Flag OE plasmid. N. IF verified pepDCBLD2-Flag was encoded and expressed in the cytoplasm in SW480 cells only transfected with circDCBLD2-Flag OE plasmid Verifying that circDCBLD2 can encode a polypeptide and promote drug resistance To validate the sequencing results, a circDCBLD2-ins-Flag fusion overexpression plasmid was constructed, as well as an overexpression plasmid circDCBLD2-ins-Flag/A-mut with a frameshift mutation caused by inserting an A in the cORF (Fig. 1J, sequence shown in Supplementary Material 6). These plasmids were transfected, and PCR was used to verify that both plasmids could upregulate the level of circDCBLD2 (Fig. 1K). Then, in the tool cell HEK293T and two colon cancer cell HCT116 and SW480, the expression of exogenous pepDCBLD2-Flag was detected using Flag-tag antibodies. WB results showed that only the wild-type circDCBLD2 could encode pepDCBLD2-Flag, while the mutant did not have the translation ability (Fig. 1L-M, Supplementary Fig. 1). By immunofluorescence (IF), it was further confirmed that only the wild-type circDCBLD2 had the encoding ability, and the polypeptide pepDCBLD2-Flag was expressed in the cytoplasm (Fig. 1N). Subsequently, wild-type or mutant plasmids overexpressing circDCBLD2 were applied in colon cancer cells, or circDCBLD2 was knocked down. An appropriate concentration of oxaliplatin (detailed in Supplementary Material 6) was added to verify that circDCBLD2 could regulate the chemoresistance of colon cancer cells by encoding pepDCBLD2. IF detection of H2AFX spot formation showed that overexpression of pepDCBLD2-Flag could weaken the DNA-damaging effect of oxaliplatin, while the circDCBLD2 mutant did not have this effect. Conversely, after knocking down circDCBLD2, due to the reduced expression of endogenous pepDCBLD2, the DNA-damaging effect of oxaliplatin on DNA was enhanced (Fig. 2A). The CCK-8 experiment confirmed that compared with the negative control group and the mutant overexpression group, overexpression of pepDCBLD2-Flag could reduce the inhibitory effect of oxaliplatin on cell proliferation. Downregulating the expression of pepDCBLD2 could significantly enhance the inhibitory effect of oxaliplatin on cell proliferation (Fig. 2B). In addition, through flow cytometry detection of cell apoptosis, it was further demonstrated that overexpression of pepDCBLD2-Flag significantly inhibited the cell apoptosis under oxaliplatin, while the apoptosis levels of the mutant group and the control group were similar. In contrast, the knockdown of pepDCBLD2 significantly promoted cell apoptosis under oxaliplatin (Fig. 2C). Verifying pepDCBLD2 could promote drug resistance. A. IF detection of H2AFX showed pepDCBLD2-Flag weakened the DNA-damaging effect of oxaliplatin. B. CCK-8 assay showed pepDCBLD2-Flag reduced the inhibitory effect of oxaliplatin on cell proliferation. C. Flow cytometry showed pepDCBLD2-Flag significantly inhibited the cell apoptosis under oxaliplatin in two groups of CRC cells Therefore, circDCBLD2 was confirmed to encode the polypeptide pepDCBLD2, and the pepDCBLD2 could enhance the resistance of colon cancer cells to oxaliplatin. Investigation on the mechanism by which pepDCBLD2 regulates the expression of its parent gene DCBLD2 DCBLD2, also known as CLCP1 or ESDN, is a transmembrane protein that contains CUB, LCCL and F5/8 type C domains [22], and participates in regulating various signal transduction pathways [23]. Previous studies have shown that high expression of DCBLD2 is associated with drug resistance in colorectal cancer [24] and poor clinical prognosis of patients, but the potential mechanism has not been reported. In this study, pepDCBLD2 was overexpressed or knocked down in two colon cancer cells. Through WB detection, it was found that overexpression of pepDCBLD2-Flag could upregulate the protein expression level of its parent gene DCBLD2, while knocking down circDCBLD2 led to a decrease in the protein level of the DCBLD2 due to reduced endogenous pepDCBLD2 expression (Fig. 3A, Supplementary Fig. 1). RT-qPCR (Supplementary Fig. 1) confirmed the knockdown efficiency of si-circDCBLD2 at the RNA level and showed that linear DCBLD2 mRNA levels remain unaffected following si-circDCBLD2 treatment. That is to say, the protein expression level of DCBLD2 is positively correlated with pepDCBLD2. Investigation on the mechanism by which pepDCBLD2 upregulates its parent gene DCBLD2. A. WB detection showed pepDCBLD2-Flag upregulated the protein expression level of its parent gene DCBLD2 in two groups of CRC cells. B. Silver staining of electrophoresis gels for Flag-IP eluate from HEK293T cells. C. Bar chart of the top 50 proteins interacted with pepDCBLD2-Flag by LC-MS/MS. D. GO annotation for pepDCBLD2-Flag-interacted proteins, among which E3 ubiquitin-protein ligase MIB1 belonged to the biological process of cell communication. E. Schematic diagram showing the competitive binding of pepDCBLD2 to MIB1 to regulate DCBLD2. F. Co-IP and WB confirmation of the interaction between pepDCBLD2-Flag and MIB1 in HEK293T, HCT116 and SW480 cells. G. IF detection of the colocalization of pepDCBLD2-Flag and MIB1 in HCT116 and SW480 cells. H. Co-IP and WB confirmation of the interaction between DCBLD2 and MIB1 in HCT116 cells. I. WB showed MIB1 OE downregulated the expression of DCBLD2 in HCT116 cells. J. WB indicated MIB1 played a crucial role in the regulatory function of pepDCBLD2 in HCT116 cells. K. IP and WB showed pepDCBLD2-Flag OE significantly inhibited the ubiquitination level of DCBLD2 in HCT116 and SW480 cells To explore the mechanism by which pepDCBLD2 regulates the expression of its parent gene, the circDCBLD2-ins-Flag fusion overexpression plasmid was transfected into the tool cell HEK293T. As shown in Fig. 3B, after electrophoresis and silver staining of the immunoprecipitation (IP) products, the gels were subjected to MS analysis, and some potential proteins interacted with pepDCBLD2-Flag were observed compared to the IgG group. Besides, MS analysis of the excised gel between 40 kDa and 55 kDa identified the specific junction-spanning polypeptides for pepDCBLD2-Flag. The top 50 interacting proteins with a differential fold greater than 2 and P-value less than 0.05 were selected and plotted in a bar chart (Fig. 3C, Supplementary Material 5). In Fig. 3D, the top 100 candidate interacting proteins was analyzed for Gene Ontology (GO) annotation using PANTHER (https://www.pantherdb.org/). Apart from the unclassified proteins, the pie chart was constructed and the sector size represented a percentage of proteins in different GO terms. It was found that most proteins were enriched in cellular processes (upper chart), and these proteins underwent further GO annotation. E3 ubiquitin-protein ligase MIB1 was enriched in cell communication (lower chart), and played a role in mediating ubiquitination and subsequent proteasomal degradation. This suggested that pepDCBLD2 was likely to compete with DCBLD2 for binding to MIB1, reducing the ubiquitination level of DCBLD2 and decreasing its degradation, thereby increasing the stability of DCBLD2 protein (Fig. 3E). To verify the hypothesis, this study performed co-IP combined with WB to confirm the interaction between pepDCBLD2-Flag and MIB1 (Fig. 3F). Through IF staining, it was proved that pepDCBLD2-Flag and MIB1 were co-localized in the cytoplasm (Fig. 3G). Additionally, MIB1 and DCBLD2 could also interact and bind to each other (Fig. 3H). Overexpression of MIB1 downregulated the expression of DCBLD2, while knockdown of MIB1 upregulated the expression of DCBLD2 (Fig. 3I, Supplementary Fig. 1). Simultaneously knocking down pepDCBLD2 and MIB1, it was found that the expression level of DCBLD2 was similar to that of the control group, indicating MIB1 played a crucial role in the regulatory function of pepDCBLD2 (Fig. 3J, Supplementary Fig. 1). Furthermore, after overexpression of pepDCBLD2-Flag in colon cancer cells, IP combined with WB detection showed that the ubiquitination level of DCBLD2 protein was significantly inhibited (Fig. 3K). In summary, pepDCBLD2 can affect the ubiquitination level of its parent gene DCBLD2 by competitively binding to MIB1, thereby increasing the protein expression of the DCBLD2. Exploration of the potential downstream signaling pathways of DCBLD2 In order to investigate the potential mechanism by which DCBLD2 regulated chemoresistance, DCBLD2 was knocked down in colon cancer cells and LC-MS/MS analysis was conducted to screen potential downstream proteins. The proteins were plotted on a volcano plot based on two factors: fold change in expression and P value (T-test). Proteins with significantly downregulated expression were labeled in blue (FC < 0.67 and p < 0.05), and proteins with significantly upregulated expression were labeled in red (FC > 1.5 and p < 0.05) (Fig. 4A). Subsequently, GO annotation was performed on the significantly different proteins. As shown in Fig. 4B, the Circos diagram was used to display the number of up-regulated and down-regulated differential proteins enriched under each GO term, the significance of GO term enrichment, the rich factor and other information. The GO terms with significant enrichment were sorted in ascending order of the P value obtained by Fisher’s exact test, and the top 20 significantly enriched GO terms were plotted as an enrichment bar chart (Fig. 4C). In addition, the differentially significant proteins were annotated with KEGG pathways, and the enrichment significance of each KEGG pathway was calculated using Fisher’s exact test. If the P value was less than 0.05, it was considered a significantly enriched pathway. As shown in Figs. 4D-E, the results of KEGG pathway enrichment analysis for up- and down-regulated differentially expressed proteins were presented in the form of a butterfly diagram and a Circos diagram. By integrating the GO annotation and KEGG pathway enrichment results, this study mainly focused on GO0004672 protein kinase activity and hsa04010 MAPK signaling pathway. The potential downstream proteins they both involve were the ERK family. Wu et al. demonstrated that DCBLD2 potentiates VEGFA-driven effects by increasing VEGFR2 phosphorylation and activating its downstream AKT and ERK1/2 signaling cascades [25]. Exploration of the Potential Downstream Signaling Pathways of DCBLD2. A. Volcano plot of the downstream proteins of DCBLD2 screened by LC-MS/MS. B. Circos diagram showing the number of up- and down-regulated differential proteins enriched under each GO term. C. Bar chart of the top 20 significantly enriched GO terms. D. Butterfly diagram of the top 10 significantly enriched KEGG pathways for differentially up- and down-regulated proteins. E. Circos diagram showing the number of up- and down-regulated differential proteins enriched under each KEGG pathway. F. STRING network predicted the regulatory relationship between ERK and PKM2. G. WB showed DCBLD2 OE activated the ERK/PKM2 pathway and upregulated the expression of PKM2. H. IF detection showed DCBLD2 OE promoted the nuclear translocation of PKM2 in HCT116 and SW480 cells. I. CCK-8 assay showed inhibition of ERK1/2 or PKM2 significantly enhanced the inhibitory effect of oxaliplatin on cell proliferation in HCT116 and SW480 cells Previous studies showed that ERK activated PKM2 and promoted glycolysis in tumor cells [26, 27]. In this study, the potential regulatory relationship between ERK and PKM2 was predicted through the STRING network (Fig. 4F). Subsequently, DCBLD2 overexpression plasmid and siRNA were used for direct overexpression or knockdown of DCBLD2. Through WB (Fig. 4G, Supplementary Fig. 2) and IF detection (Fig. 4H), it was confirmed that overexpression of DCBLD2 activated ERK1/2 and increased the level of Phospho-ERK1/2, thereby activating PKM2 and increasing the level of Phospho-PKM2, promoting the nuclear translocation of PKM2 and upregulating its protein expression. In contrast, knocking down DCBLD2 inhibited the activation of ERK1/2, thereby reducing the phosphorylation level of PKM2 and its protein expression. Besides, CCK-8 assay showed that inhibition of ERK1/2 or PKM2 reversed the chemoresistance and significantly enhanced the inhibitory effect of oxaliplatin on cell proliferation (Fig. 4I). These experiments suggested that the parent gene DCBLD2 functioned by activating the downstream ERK/PKM2 signaling pathway. Verification of the pepDCBLD2/DCBLD2/ERK/PKM2 pathway at the cellular and clinical levels After overexpression or knockdown of pepDCBLD2 in colon cancer cells, WB was performed to detect the expression levels of the ERK/PKM2 pathway. It was found that overexpression of pepDCBLD2-Flag increased the level of Phospho-ERK1/2, activated PKM2 and promoted the expression of PKM2. On the other hand, knockdown of pepDCBLD2 inhibited the activation of ERK1/2, downregulated the expression of Phospho-PKM2 and PKM2 (Fig. 5A, Supplementary Fig. 2). PKM is one of the key enzymes in glycolysis, and PKM2 is a dominant subtype of PKM in cancer cells. Thus, in this study, various kits were used to detect the related metabolic indicators. As shown in Fig. 5B and E, overexpression of pepDCBLD2-Flag enhanced glycolysis, which included increasing glucose uptake, lactate production, and ATP generation. It also enhanced the pentose phosphate pathway to increase the ratio of NADPH/NADP+, resulting in an increase in GSH levels and enhanced ROS scavenging ability. In contrast, knockdown of pepDCBLD2 inhibited glucose uptake, lactate production, and ATP generation, decreased the ratio of NADPH/NADP+, reduced GSH levels, and led to an increase in ROS accumulation intracellularly. These indicated that pepDCBLD2 could activate the DCBLD2/ERK/PKM2 signaling pathway and promote downstream glycolysis, enhancing the cellular ROS scavenging ability. Verification of the pepDCBLD2/DCBLD2/ERK/PKM2 pathway at the cellular and clinical levels. A. WB detection showed pepDCBLD2-Flag OE activated the ERK/PKM2 pathway and upregulated the expression of PKM2 in HCT116 cells. B-D. Relative glucose uptake, lactate and ATP production, relative NADPH/NADP+ level, and relative intracellular GSH level were upregulated by pepDCBLD2-Flag in HCT116 cells compared to the corresponding negative control. (n = 4, mean ± SEM, t-test). E. Flow cytometry showed intracellular ROS was downregulated by pepDCBLD2-Flag in HCT116 cells. F. TEM and WB confirmation of human serum exosomes. G. qPCR and WB detection showed the expression levels of pepDCBLD2 in the serum exosomes from drug-resistant patients (PD, n = 65) was significantly higher than those from the sensitive patients (PR, n = 59), but there was no difference in circRNA levels. H. (a) Kaplan curve of overall survival probability for patients with high or low expression of DCBLD2 using R2: Kaplan-Meier Scanner. (b) The expression graph of DCBLD2 in (a). I. IHC staining confirmed that the resistant group highly expressed the proteins related to the DCBLD2/ERK/PKM2 pathway To further validate the above signaling pathways at the clinical level, serum samples from patients with oxaliplatin-resistant colon cancer (PD, n = 65) and those with sensitive colon cancer PR, n = 59) were collected (Table 1 in Supplementary Material 6). Serum exosomes were extracted by ultracentrifugation, and their diameters of approximately 100 nm were observed under a transmission electron microscope. WB confirmed that the exosomes expressed the characteristic proteins TSG101 and CD63 (Fig. 5F). Further detection revealed that the serum exosomes of the chemoresistant patients were highly expressed with the polypeptide pepDCBLD2 (R&D Systems, AF6269), but the difference in circDCBLD2 was not significant (Fig. 5G). Searching the TCGA database (R2: Kaplan-Meier Scanner) showed that colon cancer patients with high expression of the parent gene DCBLD2 had a lower overall survival rate (Fig. 5H). Additionally, the colon cancer tissue sections were collected for immunohistochemistry (IHC) staining. It confirmed that the resistant group highly expressed the proteins related to the DCBLD2/ERK/PKM2 signaling pathway (Fig. 5I). These results indicated that the polypeptide pepDCBLD2 and its parent gene-related pathways had a carcinogenic effect. Chemoresistant cells regulated the glucose metabolism of sensitive cells through exosomal pepDCBLD2 to promote drug resistance Studies have shown that exosomes can mediate intercellular signal transduction and thereby regulate the relevant phenotypes of the receptor cells. In this study, exosomes from colon cancer oxaliplatin-resistant cells and their corresponding sensitive cells were extracted using ultracentrifugation. Transmission electron microscopy revealed a diameter of approximately 100 nm (Fig. 6A), and particle size analysis showed a roughly distribution of around 100 nm (Fig. 6B). WB confirmed that compared to the source cells, exosomes were enriched in the expression of exosome-characteristic protein TSG101 and CD63, while almost no endoplasmic reticulum-related protein Calnexin was expressed. Additionally, compared to the sensitive cells, both the resistant cells and their exosomes highly expressed the pepDCBLD2 (R&D Systems, AF6269), but there was no significant difference in circDCBLD2 (Fig. 6C). Exosome protection experiments demonstrated that pepDCBLD2 was mainly located within the exosome cavity and was protected by the membrane structure, thus avoiding degradation by proteases in the microenvironment (Fig. 6D). The exosomes of resistant cells were labeled with membrane dye PKH26 and added to the culture medium of sensitive cells for co-incubation. It was observed that sensitive cells were able to take up a large amount of exosomes from drug-resistant cells approximately 6 h later (Fig. 6E). Chemoresistant cells regulated the glucose metabolism of sensitive cells through exosomal pepDCBLD2 to promote drug resistance. A. TEM of exosomes from colon cancer cells. B. NTA of exosomes from colon cancer cells. C. qPCR detection of circDCBLD2 and WB detection of pepDCBLD2 and biomarkers in exosomes from colon cancer cells. D. Proteinase and RNase protection assays demonstrated that pepDCBLD2 was selectively resided in the exosome lumen and protected against Proteinase. E. PKH26-labeled HCT116/L-OHP exosomes could fuse into HCT116 cells. F. WB detection showed the pepDCBLD2/DCBLD2/ERK/PKM2 pathway was upregulated by exosomal pepDCBLD2 from HCT116/L-OHP cells. G-I. Relative glucose uptake, lactate and ATP production, relative NADPH/NADP+ level, and relative intracellular GSH level were upregulated by exosomal pepDCBLD2 from HCT116/L-OHP cells, compared with treatment using only PBS, or exosomes with low expression of the pepDCBLD2, or exosomes with low expression of the pepDCBLD2 but supplemented with the mutant form of circDCBLD2. (n = 4, mean ± SEM, Dunnett t-test). J. Flow cytometry showed intracellular ROS was downregulated by exosomal pepDCBLD2 from HCT116/L-OHP cells. K. CCK-8 detection showed the inhibition ratio by oxaliplatin in HCT116 cells was downregulated by exosomal pepDCBLD2 from HCT116/L-OHP cells A certain amount of exosomes from the resistant cells were added to the culture medium of the sensitive cells for co-incubation. WB analysis revealed that the exosomes from the chemoresistant cells significantly upregulated the expression of pepDCBLD2 and its parent gene DCBLD2 in the sensitive cells, and activated the downstream ERK/PKM2 pathway (Fig. 6F, Supplementary Fig. 3). Thus, these resistant exosomes enhanced glycolysis, namely increased glucose uptake, lactate production, and ATP generation (Fig. 6G), and enhanced the pentose phosphate pathway to increase the NADPH/NADP+ ratio (Fig. 6H). It increased GSH levels (Fig. 6I), enhanced the ROS scavenging ability, and reduced the accumulation of ROS induced by oxaliplatin within cells (Fig. 6J). Consequently, the exosomes from resistant cells weakened the inhibitory effect of the drug on the proliferation of colon cancer cells and promoted chemoresistance (Fig. 6K). Additionally, in colon cancer resistant cells, knocking down the expression of pepDCBLD2 or simultaneously supplementing with the mutant form of circDCBLD2, and then extracting the exosomes from these resistant cells, resulted in exosomes with low expression of the polypeptide exo sh pepDCBLD2, as well as exo sh pepDCBLD2 + ins-Flag/A-mut with low expression of the polypeptide but supplemented with the mutant form of circRNA. These two types of exosomes were also added to the culture medium of the sensitive cells for co-incubation, and it was found that neither of them could affect the expression of pepDCBLD2/DCBLD2 and its downstream pathways, the level of glucose metabolism, or the resistance to oxaliplatin within the sensitive cells (Fig. 6F and K). These results demonstrated that pepDCBLD2 was the key functional cargo responsible for transferring chemoresistance. In summary, drug-resistant exosomes transferred the polypeptide pepDCBLD2 to sensitive cells, upregulating pepDCBLD2 and its parent gene DCBLD2 in the sensitive cells, activating the ERK/PKM2 pathway, enhancing glycolysis and the pentose phosphate pathway, and promoting chemoresistance. Exploration of the role of pepDCBLD2 in promoting drug resistance of colon cancer in vivo By establishing a xenograft model in mice, the effect of pepDCBLD2 on oxaliplatin resistance in colon cancer was evaluated. As shown in Fig. 7A, HCT116 cells with stable overexpression of circDCBLD2 or the corresponding plasmid vector control were subcutaneously implanted into the groin area of all mice (n = 20). The long and short axes of each tumor were recorded every 4 days, and their volumes were calculated using the formula. When the tumors reached an appropriate size, 8 mg/kg of oxaliplatin was intraperitoneally injected every 4 days. A group of mice with stable overexpression of the mutant circDCBLD2 ins-Flag/A-mut was selected, and 20 µg of exosomes from colon cancer resistant cell HCT116/L-OHP was injected via the caudal vein every 2 days. Over time, the tumor growth in the circDCBLD2 ins-Flag overexpression group was rapid, while the tumor size of the mutant group circDCBLD2 ins-Flag/A-mut was similar to that of the negative control group, growing slowly. However, when the exosomes from chemoresistant cells were injected in the mutant group, the tumor growth significantly promoted (Fig. 7B). After 26 days, all mice were euthanized, and the tumors were removed and weighed (Fig. 7C). Compared with the control group, overexpression of circDCBLD2 ins-Flag weakened the response of the tumors to oxaliplatin, and the tumors were significantly larger and heavier. The tumors in the mutant group were similar to those in the control group, indicating that tumors in the mutant group did not have drug resistance. Nevertheless, when the exosomes from chemoresistant cells were injected into the mutant group, the tumors regained resistance, with an increased volume and weight. Exploration of the role of pepDCBLD2 in promoting drug resistance of colon cancer in vivo. A. A flowchart depicting the in vivo experimental design. B. The tumor growth in the pepDCBLD2-Flag OE group was rapid, similar to the group with both mutant circDCBLD2-ins-Flag/A-mut and exosomes from chemoresistant cells. C. Picture of the tumors in all groups and bar chart of the tumor weight. D. WB detection showed the pepDCBLD2/DCBLD2/ERK/PKM2 pathway within tumors was upregulated with pepDCBLD2-Flag OE or with both mutant circDCBLD2 and exosomes from chemoresistant cells. E. IHC staining showed the DCBLD2/ERK/PKM2 pathway and Ki67 within tumors were upregulated with pepDCBLD2-Flag OE or with both mutant circDCBLD2 and exosomes from chemoresistant cells. F. qPCR detection showed circDCBLD2 within tumors was significantly upregulated in all groups compared with the corresponding negative control. G-I. Relative lactate production, NADPH/NADP+ Ratio and GSH level within tumors were upregulated in the group with pepDCBLD2-Flag OE or with both mutant circDCBLD2 and chemoresistant exosomes, compared with the corresponding negative control. (n = 5, mean ± SEM, Dunnett t-test) Subsequently, further detection was conducted applying WB and IHC to examine the expression of pepDCBLD2/DCBLD2 and its downstream pathways within the tumors of each group. As shown in Figs. 7D (Supplementary Fig. 3) and 7E, the polypeptide pepDCBLD2-Flag and its parent gene DCBLD2 were significantly upregulated in the circDCBLD2 ins-Flag group, which subsequently activated the ERK/PKM2 pathway, and promoted cell proliferation. The mutant group did not possess the aforementioned regulatory functions. However, after adding the drug-resistant exosomes to the mutant group, the expression of pepDCBLD2/DCBLD2 and its downstream pathways was again upregulated, and the tumor proliferation activity was enhanced. Additionally, through PCR detection, it was found that the levels of circDCBLD2 in the other three groups were significantly higher compared to the control group (Fig. 7F), indicating that the activation of the DCBLD2/ERK/PKM2 pathways was independent of the circRNA itself. Due to the promoting effect of the polypeptide pepDCBLD2, the production of lactate within the tumors increased (Fig. 7G), the ratio of NADPH/NADP + and the content of GSH increased (Fig. 7H-I), thereby enhancing the ROS scavenging ability of cancer cells. Therefore, this study demonstrated that drug-resistant exosomes could transfer pepDCBLD2, and the upregulated pepDCBLD2 could increase the expression of its parent gene DCBLD2, thereby activating the ERK/PKM2 pathway, enhancing glucose metabolism, and promoting chemoresistance in vivo. Discussion To date, oxaliplatin resistance remains a major challenge in the treatment of advanced colon cancer [28]. Therefore, it is of great significance to deeply explore the mechanism of chemoresistance and potential methods to enhance the sensitivity to oxaliplatin in colon cancer. In this study, through ribosome sequencing, circDCBLD2 encoding oxaliplatin resistance-related polypeptides was screened out. It was proved that the translated polypeptide pepDCBLD2 could upregulate the protein expression of the parent gene DCBLD2. Further protein profiling analysis indicated that DCBLD2 could activate the ERK/PKM2 pathway, promote the nuclear translocation of PKM2, and upregulate PKM2 expression. Moreover, the exosomes from chemoresistant cells highly expressed pepDCBLD2 and could transfer pepDCBLD2 to sensitive cells, thereby increasing DCBLD2 expression, activating the ERK/PKM2 pathway, promoting glucose metabolism, and causing resistance in colon cancer cells (Fig. 8). This study revealed a novel mechanism by which drug-resistant exosomes carried circRNA-derived pepDCBLD2 to promote chemoresistance intercellularly in colon cancer, providing potential therapeutic targets for improving the chemotherapy efficacy of colon cancer in the future. A proposed mechanistic model in this study With the rapid development of whole-genome translation profiling technology and ribosome profiling technology, studies have found that circRNAs contain small ORFs and have potential functions in encoding polypeptides or proteins [29]. Huang et al. reported in Nature that in human breast cancer, there was a circFAM53B with unconventional translation that produced a recessive antigenic polypeptide segment binding to HLA-I, which effectively activated the initial CD4 + and CD8 + T cells in an antigen-specific manner and induced anti-tumor immune responses [18]. Besides, circSEMA3C could encode the protein SEMA3C-319aa, which upregulated the generation of polyunsaturated fatty acid chains and inhibited glutathione metabolism as well as the activity of GPX4, thereby suppressing the viability and tumor growth of gastric cancer cells in vitro and in vivo [30]. The roles of these polypeptides or microproteins in the progression of malignant tumors have been successively revealed, but it is not clear whether circRNAs can regulate the drug resistance of colon cancer by encoding polypeptides. This study applied Ribo-seq and RNA-seq to screen circDCBLD2 in colon cancer cells that was with encoding ability and related to oxaliplatin resistance. Subsequently, the ORF of circDCBLD2 was fused with the Flag tag to construct an overexpression plasmid and the corresponding insertion mutants were also constructed. Through in vitro experiments, it was further clarified that circDCBLD2 could encode the polypeptide pepDCBLD2, and pepDCBLD2 promoted the oxaliplatin resistance of colon cancer regardless of the circDCBLD2 itself. Exosomes, as a unique subgroup of extracellular vesicles, have been found to be a key mediator of intercellular communication. They transfer various molecular cargos and significantly affect the tumor microenvironment, thereby promoting the progression, metastasis, and treatment resistance of cancer [31]. Studies have shown that the molecular contents of exosomes, including miRNA, lncRNA, circRNA, etc., enable exosomes to perform functions such as drug efflux and detoxification, regulation of cell death pathways, induction of epithelial-mesenchymal transition, and inhibition of the immune system [32, 33]. However, there are few reports on the study of exosomes transporting polypeptides to regulate chemoresistance. This study clarified at the in vitro level that drug-resistant exosomes transferred the pepDCBLD2, which upregulated the expression of pepDCBLD2 and its parent gene DCBLD2 in sensitive cells, thereby promoting chemoresistance. It was also confirmed in clinical samples that the polypeptide pepDCBLD2 was highly expressed in the serum exosomes of colon cancer patients with oxaliplatin resistance compared to those of sensitive patients. Furthermore, through in vivo experiments using xenograft tumor models, it was verified that chemoresistant exosomes could transfer pepDCBLD2, and the upregulated pepDCBLD2 increased the expression of DCBLD2, promoting drug resistance in colon cancer tissues. Therefore, exploring new therapeutic strategies to overcome exosome-mediated drug resistance, such as using exosome inhibitors or designing exosome therapies, has potential application prospects. DCBLD2 was involved in regulating various signaling transduction pathways. Researchers established a prognosis model related to the Hedgehog pathway and testified that DCBLD2 was significantly upregulated in pancreatic cancer and was associated with poor prognosis [23]. Moreover, the upregulation of DCBLD2 expression in pancreatic cancer was correlated with enhanced glycolysis [34]. Additionally, high expression of DCBLD2 was associated with drug resistance in colorectal cancer and poor clinical prognosis in patients [35]. But the potential mechanism by which DCBLD2 regulates chemoresistance in colon cancer remained unclear. In this study, protein profiling combined with multiple bioinformatics analysis was performed, and it was found that DCBLD2 activated the ERK/PKM2 pathway, promoted the nuclear translocation of PKM2 and upregulated its protein expression level. This pathway enhanced glycolysis, and promoted oxaliplatin resistance in colon cancer at both in vitro and in vivo levels. Nevertheless, this study still has limitations. Due to the lack of commercialized specific antibodies, the endogenous expression of the pepDCBLD2 in colon cancer cells and clinical tissue samples was not directly detected. In this study, the antibody that recognized the 67–528 aa region of DCBLD2 (including 69–190 aa) was used for Western blot, because the pepDCBLD2 was identified as the repetitive sequence of 69–190 aa in the DCBLD2 protein. It was found that the possible endogenous pepDCBLD2 could be detected by size-based discrimination, and the observed bands were similar in position to those detected by anti-Flag. However, the possibility that the detected band represents a DCBLD2-related fragment or modified isoform cannot be fully excluded. An endogenous identification of the unique junction-spanning polypeptide is necessary for further studies. In conclusion, the circDCBLD2 could encode pepDCBLD2 through rolling circle translation. Moreover, pepDCBLD2, instead of the circRNA itself, participated in competitive binding to ubiquitin protein ligases, thereby upregulating the expression of the parent gene DCBLD2 and promoting chemoresistance in colon cancer. Additionally, exosomes from drug resistant cells transferred pepDCBLD2 to sensitive cells, intercellularly upregulating DCBLD2 and activating the downstream ERK/PKM2 signaling pathway, thereby increasing glucose metabolism and ROS scavenging capacity, and leading to chemoresistance. This provided a novel potential target for the treatment of oxaliplatin-resistant colon cancer in the future. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Supplementary Fig. 1. Quantification of WB images in Figs. 1 and 3 by grayscale analysis (n = 3, mean ± SEM, t-test or Dunnett t-test), and qPCR detection of circDCBLD2 and DCBLD2 mRNA with si-circDCBLD2 or si-DCBLD2 Supplementary Material 2: Supplementary Fig. 2. Quantification of WB images in Figs. 4 and 5 by grayscale analysis (n = 3, mean ± SEM, t-test) Supplementary Material 3: Supplementary Fig. 3. Quantification of WB images in Figs. 6 and 7 by grayscale analysis (n = 3, mean ± SEM, Dunnett t-test) Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Supplementary Material 7 Acknowledgements Not applicable. Author contributions XW, JY, HY and YX performed all of the experiments. XW designed the experiments, analyzed the data and wrote the manuscript. HW and MS analyzed the data and edited the manuscript. HZ, YW and MG supported the experimental platform and reviewed the manuscript. HZ is the guarantor of this work, has full access to all of the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis. Funding This work was supported by grants from the National Natural Science Foundation of China (Nos. 82303924), and the Tianjin Clinical Key Specialty Construction Project – Malignant Tumors of the Digestive System (No. 2024ZDZK017). The funders had no role in the study design, the data collection and analysis, the interpretation of the data, the writing of the article, and the decision to submit this article for publication. Data availability The datasets generated and/or analyzed during the current study are not publicly available because further in-depth research is still being carried out within our team, but are available from the corresponding author on reasonable request. Declarations Ethical approval All animal experiments were performed in line with protocols approved by the Institutional Animal Care and Research Advisory Committee of Nankai University (2023-SYDWLL-000113). Guide for the Care and Use of Laboratory Animals was followed in the study. The ethical approval document was uploaded to the attachment. All studies involving human samples were performed in line with protocols approved by the Science and Technology Ethics Committee of Tianjin First Central Hospital (Research 20241120-1). Consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Ming Gao, Email: headandneck2008@126.com. Ye Wang, Email: wangye950@163.com. Haiyang Zhang, Email: zhanghaiyang@tmu.edu.cn. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are not publicly available because further in-depth research is still being carried out within our team, but are available from the corresponding author on reasonable request.