Extracellular vesicle-mediated delivery of circp53 suppresses the progression of multiple cancers by activating the CypD/TRAP/HSP90 pathway

Abstract

The majority of cancers remain incurable due to limited therapeutic responses in malignancies with high-risk genetic mutations such as TP53. Building on the success of mRNA vaccine technology, we investigated circular RNA (circRNA) therapeutics and identified hsa_circp53_0041947, a TP53-derived circRNA in multiple myeloma (MM). The hsa_circp53_0041947 encodes a functional peptide (circp53-209aa) demonstrating p53 mutation-independent anti-MM effects through CypD/TRAP1/HSP90 complex-mediated mechanisms. Specifically, circp53-209aa activated cyclophilin D (CypD) isomerase activity at the circp53-209aa-R175 site, triggering mitochondrial permeability transition pore opening and subsequent mitochondrial apoptosis. To enable targeted delivery, we engineered extracellular vesicle (EV) systems, E7-Lamp2b-EVs and Her2-Lamp2b-EVs, for MM and colorectal cancer, respectively. Circp53-EVs administration achieved tumor-selective growth inhibition in both malignancies. Our study establishes engineered circp53-EVs as a versatile therapeutic platform, demonstrating the translational potential of circRNA-based strategies for refractory cancers with TP53 pathway alterations.

Fig. 1. Circp53 expression is decreased in patients with MM and is associated with superior outcomes.

a, An illustration of the annotated genomic region of TP53, the putative different RNA splicing forms and the validation strategy for circular exons 5 and 11 (circp53). Convergent (blue) and divergent (red) primers were designed to amplify the linear or back-splicing products. b, RNA levels of circp53 and linear TP53 were determined by PCR with and without RNase R treatment. c, RNA levels of circp53 and linear TP53 were determined by RT–qPCR with and without RNase R treatment. d, Sanger sequencing following PCR was conducted using the indicated divergent flanking primers, confirming the ‘head-to-tail’ splicing of circp53 in MM1.S, H929, XG1 and RPMI 8226 cells. e, The putative ORF in circp53. The sequences of the putative ORF are shown in green. f, The predicted sequence of circp53–209aa. g, WB analysis of endogenous circp53–209aa expression in MM1.S, H929, XG1 and RPMI 8226 cells. h, The specific peptides from circp53–209aa were identified by MS analysis. i, The levels of circp53 in patients with MM were lower than in those in NP as evaluated by BaseScope analysis and the representative staining images are shown with positive reactions indicated by red arrows. j, Statistical analysis of BaseScope analysis (n = 6 clinical samples for each group, P < 0.001). k, Circp53 levels were significantly decreased in patients with MM (n = 48 samples for each group, P < 0.001). l, Higher levels of circp53 were associated with longer overall survival (OS) survival. The data are presented as mean ± s.d. ^*P < 0.05, ^**P < 0.01 and ^***P < 0.001.

Fig. 2. Circp53 activates the mitochondrial apoptotic pathway.

a, An illustration of endo-circp53 and circp53-HA. b, RT-qPCR was used to determine the RNA levels of circp53 and linear TP53. c, PCR was used to determine the RNA levels of circp53 and TP53. d, Sanger sequencing was performed following PCR using the indicated divergent primers to confirm the precise splicing of the circp53-OE plasmid in MM1.S, H929, XG1, and RPMI 8226 cells. The circRNA hsa_circ_0000284 was used as an internal control gene. e, WB and MS analyses of circp53–209aa overexpression in MM1.S, H929, XG1 and RPMI 8226 cells using the HA-tag antibody. f, Confocal microscope live images showing the cellular localization of circp53–209aa. g, The MTT assay demonstrated decreased cell proliferation in circp53-OE cells compared with Ctrl cells (P < 0.01). h, The EdU incorporation assay revealed a significant decrease in the number of proliferating cells in circp53–209aa-OE cells compared with Ctrl cells. i, Statistical analysis of EdU incorporation assay. j, The TUNEL assay showed a significant increase in the number of apoptotic cells in circp53–209aa-OE cells compared with Ctrl cells. k, Statistical analysis of EdU incorporation assay. l, The scatterplot of KEGG pathway enrichment analysis for RNA-sequencing data revealed that the molecular function of circp53–209aa was centered on the p53 signaling pathway. m, The par chart of KEGG pathway enrichment analysis for RNA-sequencing data also revealed that the molecular function of circp53-209aa was centered on the p53 signaling pathway. n, The KEGG enrichment analysis revealed that the p53 signaling pathway was enriched, ranking within the top 20 pathways. o, WB analysis demonstrated that increased levels of circp53 induced apoptosis, as indicated by the cleavage of the apoptotic regulators PARP and Caspase 3. p, Statistical analysis of the expression of cleavage of the apoptotic regulators PARP and Caspase 3. q, WB analysis determined the expression levels of the mitochondrial apoptotic pathway associated proteins Bad, Bak, Bcl-xL, Bax and Bcl-2 in circp53-OE cells. r, Statistical analysis of the expression of the mitochondrial apoptotic pathway associated proteins. s, Images of CDX model mice on day 28. t, Tumors taken from the CDX model in each group. u, The tumor growth curve of the model mice in the Ctrl and circp53-OE groups. v, The tumor weight of the CDX model mice in the Ctrl and circp53-OE groups. The data are presented as mean ± s.d. (n = 6 mice for each group, n = 3 cultures for each group, ^*P < 0.05, ^**P < 0.01 and ^***P < 0.001).

Fig. 3. Combining circP53–209aa with BTZ synergistically inhibits the proliferation of MM cells by activating the mitochondrial apoptotic pathway.

a, The effects of BTZ on cell apoptosis in MM cells with or without circp53 overexpression. The circp53-OE cells were more sensitive to BTZ treatment compared with WT cells (P < 0.01). b, Statistical analysis of cell apoptosis. c, Representative images from the TUNEL assay demonstrate a significant increase in the number of apoptotic cells in the circp53-OE group compared with the Ctrl group after BTZ treatment (n = 3 cultures for each group, P < 0.01). d, Statistical analysis of TUNEL assay. e, Representative images from the EdU incorporation assay show a significant decrease in the number of proliferating cells in the circp53-OE group compared with the Ctrl group upon BTZ treatment (n = 3 cultures for each group, P < 0.01). f, Statistical analysis of EdU incorporation assay. g, Representative images from the soft agar plates demonstrate a significant decrease in colony-forming efficiency in the circp53-OE group compared with the Ctrl group upon BTZ treatment. h, Statistical analysis of colony-forming efficiency (n = 3 cultures for each group, P < 0.01). i, The MTT assay results showed that elevated levels of circp53–209aa significantly increased the sensitivity of p53 mutant cells (XG1 and RPMI 8226) to BTZ treatment. j, The MTT assay results showed that elevated levels of circp53–209aa significantly increased the sensitivity of p53 WT cells (MM1.S and H929) to BTZ treatment. k, Treatment with EVs overcame BTZ resistance in p53 mutant cells such as XG1 and RPMI 8226 cells. l, Treatment with EVs overcame BTZ resistance in p53 WT cells such as H929 and MM1.S cells. m, Photographic images of PDX model mice. n, Photographic images of PDX model tumors. o, The combination of circp53-EVs and BTZ significantly inhibited mean tumor volume. p, The combination of circp53-EVs and BTZ significantly inhibited mean tumor weight (P < 0.001). The data are presented as mean ± s.d. (n = 5 mice for each group, n = 3 cultures for each group, *P < 0.05, **P < 0.01 and ***P < 0.001). NC, normal control.

Fig. 4. Circp53–209aa-R175 is identified as the site-specific target of CypD to trigger mPTP opening in MM cells.

a, GO analysis of RNA-sequencing data revealed that the circp53–209aa is highly associated with mitochondrial pathways. b, Go analysis of RNA-sequencing data revealed that the circp53–209aa is highly associated with apoptotic process signaling pathways. c, A graphic illustration of circp53–209aa interacting with CypD and releasing CypD from the CypD/TRAP1/HSP90 complex, thereby activating its isomerase activity and inducing mPTP opening. d, A Co-IP assay confirmed the direct interaction between circp53–209aa and CypD in MM cells. e,Representative confocal images of HA and CypD demonstrate the interaction between circp53–209aa and CypD. f, Statistical analysis of the colocation of circp53-209aa and CypD in H929 and XG1 cells. g, A Co-IP assay showed a weaker interaction between CypD and HSP90 in circp53-OE MM cells compared with Ctrl cells. h, Distribution of calcein in H929 and MM1.S cells. i, Distribution of calcein in XG1 and RPMI 8226 cells. j, IOD quantitative statistics of Calcein in Ctrl and circp53–209aa-OE MM cells. k, ATP levels were significantly decreased in circp53–209aa-OE cells. l, Statistical analysis of ATP levels in different cells (P < 0.001). m, A model showing the high confidence interval of the CypD protein. n, A model of the circp53–209aa-CypD complex. R175 of circp53–209aa binds to E43, E34 and Q163 of CypD through hydrogen bonds and/or ionic interactions. o, Circp53–209aa-R175 is involved in hydrogen-bond interactions with CypD, and these interactions are disrupted upon mutation of arginine to a histidine, as demonstrated by Co-IP using an HA antibody as bait in circp53-OE cells, followed by WB analysis. p, Circp53–209aa-R175H-OE MM cells were constructed successfully, as shown by WB analysis. q, Circp53–209aa-R175H did not influence the proliferation of MM cells. r, Circp53–209aa interacts with CypD and releases it from the CypD/TRAP1/HSP90 complex, as evidenced by Co-IP. s, Colocalization dynamics of CypD/HSP90 in H929 and MM1.S cells. t, Colocalization dynamics of CypD/HSP90 in XG1 and 8226 cells. u, Statistical analysis of the colocation of CypD and HSP90 in H929 and MM1.S cells. v, Statistical analysis of the colocation of CypD and HSP90 in XG1 and RPMI 8226 cells. w, IOD quantitative statistics of CypD and HSP90 in MM cells. x, Distribution of calcein in circp53-OE and circp53–209aa-R175 MM cells. y, The mPTP was not opened in the circp53–209aa-R175H group by quantitative statistics. The data are presented as mean ± s.d. (n = 3 cultures for each group, *P < 0.05, **P < 0.01, ***P < 0.001 and ns, no statistical significance).

Fig. 5. Circp53–209aa activates the mitochondrial apoptosis pathway in colorectal, lung, stomach, liver and breast cancer cells.

a, The RNA levels of circp53 and linear p53 were determined by RT–PCR with and without RNase R treatment in RKO, HCT116, HepG2, SGC-7901, A549 and MCF-7 cells. b, The RNA levels of circp53 and linear p53 were determined by RT–qPCR with and without RNase R treatment in RKO, HCT116, HepG2, SGC-7901, A549 and MCF-7 cells. c, The levels of circp53 were significantly lower in patients with CRC compared with NP, as evaluated by RNAscope analysis. Representative staining images are shown, with positive reactions indicated by red arrows. d, Statistical analysis of RNAscope analysis (P < 0.01) .e, RNA levels of circp53 were determined by RT–PCR. f, WB analysis was performed to examine overexpression of circp53–209aa in RKO, HCT116, HepG2, SGC-7901, A549 and MCF-7 cells using the HA-tag antibody. g, The specific peptides from circp53–209aa were identified by MS analysis. h The MTT assay demonstrated decreased cell proliferation rates of circp53-OE RKO, HCT116 cells compared with Ctrl cells. i, The MTT assay demonstrated decreased cell proliferation rates of circp53-OE HepG2, SGC-7901 cells compared with Ctrl cells. j, The MTT assay demonstrated decreased cell proliferation rates of circp53-OE A549 and MCF-7 cells compared with Ctrl cells. k The EdU incorporation assay demonstrated the numbers of proliferating cells were significantly decreased in circp53–209aa-OE cells compared with Ctrl cells. l, Statistical analysis of EdU incorporation assay. m, The effects of circp53 overexpression on cell apoptosis in RKO, HCT116, HepG2, SGC-7901, A549 and MCF-7 cells. n, Statistical analysis of cell apoptosis. o, The TUNEL assay showed the numbers of apoptotic cells were significantly increased in circp53–209aa-OE cells compared with Ctrl cells. p, Statistical analysis of TUNEL assay. q, Flow cytometry showed an increased level of circp53–209aa, resulting in a decrease in the ATP level in cancer cells. r, Confocal microscopic analysis revealed a rapid and almost complete loss of Calcein fluorescence signal in A549, HepG2, SGC-7901 and MCF-7 circp53–209aa-OE cells. s, Statistical analysis of confocal microscopic analysis in A549, HepG2, SGC-7901 and MCF-7 circp53–209aa-OE cells. t, Confocal microscopic analysis revealed a rapid and almost complete loss of Calcein fluorescence signal in RKO and HCT116 circp53–209aa-OE cells. u, Statistical analysis of confocal microscopic analysis in RKO and HCT116 circp53–209aa-OE cells. v, The effects of circp53 overexpression on the mitochondrial apoptotic pathway associated proteins Bad, Bak, Bcl-xL, Bax and Bcl-2 in RKO and HCT116 cells. w, The effects of circp53 overexpression on the mitochondrial apoptotic pathway associated proteins Bcl-xL and Bcl-2 in HepG2 and SGC-7901 cells. x, Statistical analysis of expression of mitochondrial apoptotic pathway associated proteins Bad, Bak, Bcl-xL, Bax and Bcl-2 in RKO and HCT116 cells. y, Statistical analysis of expression of mitochondrial apoptotic pathway associated proteins Bad, Bak, Bcl-xL, Bax and Bcl-2 in HepG2 and SGC-7901 cells. The data are presented as mean ± s.d. (n = 6 clinical samples for each group, n = 3 cultures for each group, *P < 0.05, **P < 0.01 and ***P < 0.001).

Fig. 6. Construction of the E7-circp53-EV and Her2-circp53-EV delivery systems specifically targeting BMSCs and CRC cells, respectively.

a, A schematic representation of the production and collection process of E7-circp53-EVs and Her2-circp53-EVs for targeted circp53 delivery. b, Total RNA extracted from Her2-circp53-EVs was incubated with or without RNase R, followed by RT–qPCR analysis. c, Total RNA extracted from E7-circp53-EVs was incubated with or without RNase R, followed by RT–qPCR analysis. RNase R treatment decreased the linear RNA level of GAPDH, but did not affect circp53. d, Laser-scanning confocal microscopy was used to record the uptake of PKH26-labeled EVs (red) by RPMI 8226 cells (green). e, Statistical analysis of Laser-scanning confocal microscopy of RPMI 8226 cells. f, Laser-scanning confocal microscopy was used to record the uptake of PKH26-labeled EVs (red) by H929 cells (green). g, Statistical analysis of Laser-scanning confocal microscopy of H929 cells. h, Laser-scanning confocal microscopy was used to record the greater uptake of E7-circp53-EVs compared with circp53-EVs in BMSCs, demonstrating a high specific affinity of the E7 peptide sequence to BMSCs. i, Statistical analysis of Laser-scanning confocal microscopy of BMSCs. j, Laser-scanning confocal microscopy was used to record the greater uptake of E7-circp53-EVs in BMSC cells compared with RKO, HepG2, HT29, H929 and LO2 cells. k, Statistical analysis of Laser-scanning confocal microscopy of RKO, HepG2, HT29, H929 and LO2 cells. l, Laser-scanning confocal microscopy was used to record the greater uptake of Her2-circp53-EVs compared with circp53-EVs in RKO, confirming the ability of Her2-circp53-EVs to target CRC cells. m, Laser-scanning confocal microscopy was used to record the greater uptake of Her2-circp53-EVs compared with circp53-EVs in HT29 cells. n, Statistical analysis of Laser-scanning confocal microscopy of RKO and HT29 cells. o, Laser-scanning confocal microscopy was used to record the greater uptake of Her2-circp53-EVs in RKO cells compared with H929, LO2 and HCOEPIC cells. p, Statistical analysis of Laser-scanning confocal microscopy of EVs marked by PKH26. The data are presented as mean ± s.d. (n = 3 cultures for each group, *P < 0.05, **P < 0.01 and ***P < 0.001).

Fig. 7. E7-circp53-EVs and Her2-circp53-EVs selectively inhibit tumor growth in PDX and NOD/SCID-TIBIA mouse models.

a, An IVIS spectrum was recorded to show the high fluorescence intensity in the spine and prolonged retention time in the E7-circp53-EVs group, further confirming its ability to target bone in vivo. b, Images of NOD/SCID-TIBIA mice in NC, circp53-EVs and E7-circp53-EVs groups. c, Human κFLC levels in mouse serum were measured by an ELISA. d, Representative micro-computed tomography images of bones in the NC, circp53-EVs and E7-circp53-EVs groups. e, The BMD of NOD/SCID-TIBIA mice in the NC, circp53-EVs, and E7-circp53-EVs groups. f, The BV/TV of NOD/SCID-TIBIA mice in the NC, circp53-EVs and E7-circp53-EVs groups. g, h, Images of PDX model mice on day 27 (g) and tumors taken from mice in each group (h) (P < 0.001). i, Tumor growth curves of the PDX model for the NC, circp53-EVs and Her2-circp53-EVs groups. j, Tumor weights of the PDX model in the NC, circp53-EVs and E7-circp53-EVs groups. k, IVIS was employed to observe the fluorescence in the NC, circp53-EVs and E7-circp53-EVs groups. l, RT–qPCR analysis revealed that Her2-circp53-EVs accumulated to a greater extent in PDX tumors compared with other organs such as the heart, liver, spleen, lung and kidney. m, Confocal imaging of EVs biodistribution in murine organs. n, Confocal microscopy of frozen sections showed that Her2-circp53-EVs accumulated to a greater extent in PDX tumors compared with other organs such as the heart, liver, spleen, lung and kidney by quantitative statistics. The data are presented as mean ± s.d. (n = 6 mice for each group, ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 and ns, no statistical significance).

Fig. 8. A schematic diagram illustrates how the targeted delivery of circp53 via EVs suppresses cancer progression by opening the mPTP.

a, The process of packaging tumor-targeted EVs that overexpress circp53. b, The mechanism of tumor cell cytotoxicity mediated by EVs that overexpress circp53. c, The targeted delivery of circp53-enriched EVs to CRC and MM.