Reprogramming the tumor microenvironment with c-MYC-based gene circuit platform to enhance specific cancer immunotherapy

Abstract

Intratumor heterogeneity (ITH) is associated with anti-tumoral immune response and with the efficiency of cancer immunotherapy, yet overcoming ITH remains a significant challenge. Notably, cellular MYC (c-MYC) has been shown to be a pivotal orchestrator of this ITH progression. Here, we develop a c-MYC-based sensing circuit (cMSC) that is activated exclusively by aberrant c-MYC levels, along with an exosome-based cell-to-cell (CtC) system that augments communication among tumor cells, effectively targeting all cells in tumors circumventing the limitations imposed by ITH. Further expression of multifunctional immunostimulatory agents in these cMSC-reprogrammed cancer cells remodels the tumor microenvironment, enhancing selective T-cell-mediated oncolysis. Our cMSC/CtC platform specifically senses aberrant c-MYC expression and subsequently triggers a robust cancer immunotherapeutic response. These findings offer a promising avenue for targeting cancers via precisely sensing c-MYC, overcoming the limitations of ITH.

Fig. 1. Construction of a c-MYC-based gene circuit for reprogramming cancer cells to trigger specific cancer immunotherapies.

Overall design of the gene circuit platform. An intravesical instillation therapy based on the cMSC/CtC platform has been delivered by AAV and developed to eradicate bladder cancer. The cMSC is programmed to be activated precisely in cancer cells with pathological c-MYC expression. The cell-to-cell system is designed to augment gene circuit communication in tumor cells. Combinatorial immunostimulatory outputs, including STE bitargeting CD3 and CD28, IL-21, CCL5, and anti-PD1, are expressed by the cMSC/CtC platform to selectively trigger robust cancer immunotherapy. Created in BioRender. Zhan, H. (2025) https://BioRender.com/3egw3g5.

Fig. 2. Design and implementation of the cMSC.

a Schematic showing synthetic P_aMYC expressing GFP in MYC^low and MYC^high cells. P_aMYC exhibited low transcriptional activity in MYC^low cells but high activity in MYC^high cells. b Schematic illustration of synthetic P_rMYC expressing mCherry across c-MYC expression gradients. P_rMYC exhibited high transcriptional activity in MYC^low cells but low activity in MYC^high cells. c Mean fluorescence intensity (MFI) of GFP- and mCherry-driven expression of P_aMYC and P_rMYC in BFP⁺ cells, respectively, over a range of siRNA and tetracycline concentrations. The data are presented as the mean ± SD, n = 3 individual experiments. d Schematic showing the composition of the preliminary cMSC and its working mechanism in MYC^low and MYC^high cells. The preliminary cMSC consisted of two parts. The first part was P_aMYC, which expresses the GOI expression cassette whose 3’-UTR contains ten sequences of split-ribozyme-1 (split-1) in tandem. The second part consisted of P_rMYC expressing split-ribozyme-2 (split-2). The working mechanisms of the preliminary cMSC in cells are as follows: In MYC^low cells (left), low expression of c-MYC results in low activity of P_aMYC and high activity of P_rMYC. In step 1, P_aMYC resulted in weak mRNA (GFP-10×split-1) transcription, whereas P_rMYC exhibited strong transcription of split-2. In step 2, a large amount of split-2 would interact with GFP-split-1, leading to the degradation of GFP mRNA through the ribozyme-based mRNA degradation process, resulting in almost no expression of GFP (step 3). In MYC^high cells (right), high expression of c-MYC resulted in high activity of P_aMYC and low activity of P_rMYC. P_aMYC resulted in strong mRNA (GFP-10×split-1) transcription, whereas P_rMYC resulted in weak transcription of split-2. In step 2, a small amount of split-2 would partly interact with GFP-split-1, leading to only partial degradation of the GFP mRNA, which led to a small amount of GFP translation (step 3). e Schematic showing the composition of the optimized cMSC and its working mechanism in MYC^low and MYC^high cells. The optimized cMSC consists of two parts. First, 6 binding sites of KLFa (6×KLFa-BS) were inserted upstream of P_aMYC. The reengineered P_aMYC (containing KLFa-BS) was used to drive the expression of a 5’ cassette encoding GFP, a T2A coding sequence, a cassette encoding KLFa and 3’ split-ribozyme-1 in tandem (GFP-T2A-KLFa-10×split-1). In the other part, P_rMYC was used to express split-ribozyme-2 (split-2). The working mechanisms of the optimized cMSC in cells are as follows: In MYC^low cells (left), low expression of c-MYC results in low activity of P_aMYC and high activity of P_rMYC. In step 1, P_aMYC resulted in weak mRNA (GFP-T2A-KLFa-10×split-1) transcription, whereas P_rMYC exhibited strong transcription of split-2. In step 2, a large amount of split-2 would interact with the mRNA of GFP-T2A-KLFa-10×split-1, leading to the degradation of GFP-T2A-KLFa-10×split-1 mRNA through the ribozyme-based mRNA degradation process, resulting in almost no expression of GFP (step 3). In MYC^high cells, high expression of c-MYC resulted in high activity of P_aMYC and low activity of P_rMYC. In step 1, a large amount of GFP-T2A-KLFa-10×split-1 is transcribed by P_aMYC, whereas a small amount of split-2 is produced by P_rMYC. In step 2, some GFP-T2A-KLFa-10×split-1 mRNAs interacted with split-2 and were degraded, whereas others were retained for the next step of translation. In step 3, small amounts of GFP and KLFa were translated. In step 4, the successfully translated KLFa is transferred to the nucleus, where it binds to the KLFa-BS upstream of P_aMYC, reactivating it (step 5) to transcribe more GFP-T2A-KLFa-10×split-1 mRNA (step 6) and translation (step 7). This positive feedback loop would generate an increasing amount of GFP-T2A-KLFa-10×split-1 mRNAs so that the degradation effect of split-2 transcribed by P_rMYC on GFP-T2A-KLFa-10×split-1 mRNA was negligible in this scenario. f Determining the mean fluorescence intensity (MFI) of GFP-driven expression by the preliminary cMSC and the optimized cMSC in BFP⁺ cells over a range of siRNA and tetracycline concentrations. The data were presented as the mean ± SD, n = 3 individual experiments.

Fig. 3. Design and characterization of the CtC platform.

a Schematic showing the proof-of-principle assay of the CtC platform. Two expression vectors were used here. First, there were two independent cMSC systems expressing GFP^DR (cMSC-GFP^DR) and mCherry (cMSC-mCherry). The cMSC-GFP^DR and cMSC-mCherry strains were constructed with the same expression vector (Reporter vector). Second, a cMSC was used to drive the expression of the 5’ CD63-trCas13d expression cassette and the T2A and 3’ MUC1 scFv-Lamp2b expression cassettes (CtC expression vector). In the cells transfected with the reporter vector, both GFP and mCherry were expressed in the MYC^high cells simultaneously. In the cells cotransfected with the reporter vector and the CtC expression vector, GFP was expressed in the MYC^high cells, and its mRNA could be packaged into the CtC exosomes for shuttling to nearby cells, especially those with high MUC1 expression. mCherry was expressed in only the MYC^high cells, which were used to label the transfected cells. b–d Representative images (b) of cells transfected with the reporter vector alone (mock group, top) and cotransfected with the reporter vector and the CtC expression vector (CtC group, bottom) are shown. FACS analysis of GFP and mCherry expression (c). FACS was used to quantify the proportion of GFP-positive and mCherry-positive cells among the 293T cells transfected as described above (d). These data indicated that the presence of CtC events resulted in more cells in the group expressing GFP. The data were presented above as the mean ± SD, n = 3 individual experiments, and significance was determined via two-tailed Student’s t-tests, NS not significant. Scale bar: 300 μm (the line in white) and 50 μm (the line in red). e Quantification of the expression levels of GFP and mCherry in a qPCR assay. These data indicated that the presence of CtC events could not increase the total amount of GFP expressed in a cell population. The data were presented above as the mean ± SD, n = 3 individual experiments, and significance was determined via two-tailed Student’s t-tests; NS not significant. f, g Illustration of CtC events in a cancer xenograft model. GFP and mCherry expression in T24 tumor tissue infected with AAV (cMSC-GFP^DR-cMSC-mCherry) alone or with both AAV (cMSC-CD63-dCas13d) and AAV (MSC-GFP^DR-cMSC-mCherry) (f). Areas marked by white arrows represent spatial segregation of GFP and mCherry fluorescence, indicating nonoverlapping cellular expression domains. FACS quantification confirmed the differential distribution of GFP-positive versus RFP-positive cells (g). The data were presented above as the mean ± SD, n = 3 individual experiments, and significance was determined via two-tailed Student’s t-tests. Scale bar: 300 μm (the line in white) and 50 μm (the line in red). h, i Illustration of CtC events in bladder cancer organoids. Quantification of GFP-positive and mCherry-positive cells in bladder cancer organoids (h). The tissue indicated by the white arrow is the tissue where GFP and mCherry expression are not colocalized. Bladder cancer organoids were infected with the above treatments (i). The data were presented above as the mean ± SD, n = 3 individual experiments, and significance was determined via two-tailed Student’s t-tests. Scale bar: 300 μm (the line in white) and 50 μm (the line in red).

Fig. 4. Improving the efficiency of cancer immunotherapy via the cMSC/CtC platform.

a Schematic showing cMSC/CtC-mediated STE diffusion in cell populations to improve the efficiency of cancer immunotherapy. In panel 1, CtC events were not detected. In a tumor population, only some cancer cells expressed STE because of ITH or incomplete transfection efficiency, which resulted in low immunotherapy efficiency. In panel 2, CtC events were present. cMSC/CtC-mediated STE diffusion in cell populations increases the number of STE-expressing cancer cells, which results in improved efficiency of cancer immunotherapy. b Mock group: T24 cells transfected with the cMSC vector expressing a nonfunctional polypeptide; cMSC group: T24 cells transfected with the cMSC vectors expressing STE for targeting CD3 and CD28 (cMSC-STE^DR); cMSC/CtC^S group: T24 cells cotransfected with cMSC vectors expressing STE for targeting CD3 and CD28 (cMSC-STE^DR) and a cMSC vector expressing the CtC expression component (cMSC-CD63-trCas13d-T2A-MUC1 scFv-Lamp2b). To characterize STE display and expression levels on cell surfaces, the transfected T24 cells were stained with phycoerythrin-labeled anti-His tag Ab/PE. The proportion of STE-expressing cells in each group was evaluated by flow cytometry. c Proportions of dead T24 cells in the mock, cMSC and cMSC/CtC^S groups; n = 3 individual experiments. The data were presented above as the mean ± SD, and significance was determined via two-tailed Student’s t-tests. d–f The expression levels of TNF-α (d), IL-2 (e), and IFN-γ (f) in the medium after incubation with PBMCs from T24 cells in the mock, cMSC and cMSC/CtC^S groups, respectively, n = 3 individual experiments. The data were presented above as the mean ± SD, and significance was determined via two-tailed Student’s t-tests. g CtC-mediated STE diffusion can further improve the efficiency of immunotherapy induced by a combination of immunostimulatory factors. In panel 1, CtC events were not detected. In a tumor population, only some cancer cells expressed STE because of ITH or incomplete transfection efficiency, which resulted in low immunotherapy efficiency in non-STE-expressing cancer cells. In panel 2, CtC events were present. cMSC/CtC-mediated STE diffusion in cell populations increases the number of STE-expressing cancer cells, which results in high-efficiency cancer immunotherapy. h Schematic illustration of the treatment schedule in the mouse model of T24 bladder cancer cells. Created in BioRender. Zhan, H. (2025) https://BioRender.com/ckseent. i, j Average tumor weight (i) and average tumor growth kinetics (j) of the mice subjected to different treatments: mock, cMSC, cMSC^SIPC, and cMSC/CtC^S. The tumor volume (mm³) was monitored daily using a caliper. The data were presented as the mean ± SD (n = 5 mice), and significance was determined via two-tailed one-way ANOVA. k Kaplan–Meier survival curves of various groups. A log-rank (Mantel‒Cox) test was performed to compare survival between groups (n = 5 mice per group). l The level of serum IFN-γ was determined by ELISA on Days 0, 7, 14, and 21. The data were presented as the mean ± SD (n = 5 mice), and significance was determined via two-tailed Student’s t-tests. m Mouse body weight was monitored and recorded. The data were presented as the mean ± SD (n = 5 mice), and significance was determined via two-tailed one-way ANOVA; NS not significant.

Fig. 5. cMSC^SIPC/CtC^S triggers potent T-cell-mediated tumor-specific killing without the limitations of ITH.

a–c Percentages of dead RT4-GFP (a), RT4-mCherry (b), and RT4-GFP/RT4-mCherry (c) cells treated with PBS, cMSC^SIPC or cMSC^SIPC/CtC^S, respectively. The data were presented as the mean ± SD; n = 3 individual experiments. Significance was determined via two-tailed Student’s t-tests; NS not significant. d–f Total expression levels of TNF-α (d), IL-2 (e), and IFN-γ (f) in the supernatants of cells in the PBS, cMSC^SIPC, and cMSC^SIPC/CtC^S groups, respectively, n = 3 individual experiments. The data were presented as the mean ± SD, and significance was determined via two-tailed Student’s t-tests. NS not significant. g Schematic illustration of the treatment schedule for the RT4-GFP/RT4-mCherry mouse model. Created in BioRender. Zhan, H. (2025) https://BioRender.com/ckseent. h Comparison of the tumor growth kinetics of RT4-GFP/RT4-mCherry mice subjected to different treatments: PBS, AAV-blank, anti-PD1, AAV (cMSC^SIPC), and AAV (cMSC^SIPC/CtC^S). The tumor volume (mm³) was monitored daily using the caliper method. The data were presented as the mean ± SD (n = 5 mice), and significance was determined via two-tailed one-way ANOVA. NS not significant. i Kaplan‒Meier survival curves of various groups. A log-rank (Mantel‒Cox) test was performed to compare survival between groups (n = 5 mice). NS not significant. j, k Percentages of CD69 + T cells in the CD4+ (j) and CD8+ (k) T-cell subsets among tumor-infiltrating lymphocytes from different groups (n = 5 mice). The data were presented as the mean ± SD; significance was determined via two-tailed Student’s t-tests. NS not significant. l, m Proportions of IFN-γ-secreting CD4+ (l) and CD8+ (m) T cells in the tumor microenvironments of the different groups (n = 5 mice). The data were presented as the mean ± SD; significance was determined via two-tailed Student’s t-tests. NS not significant. n, o Frequencies of TNF-α-producing CD4+ (n) and CD8+ (o) T lymphocytes in tumors from different groups (n = 5 mice). The data were presented as the mean ± SD; significance was determined via two-tailed Student’s t-tests. NS not significant. p Schematic illustration of the treatment schedule in the mouse model of PDX. q The average tumor growth kinetics of PDXs demonstrated different responses to various treatments: PBS, AAV-blank, anti-PD1, AAV (cMSC^SIPC), and AAV (cMSC^SIPC/CtC^S). The tumor volume (mm³) was monitored daily using a caliper. The data were presented as the mean ± SD (n = 5 mice), and significance was determined via two-tailed one-way ANOVA. NS not significant. r Kaplan‒Meier survival curves of various groups. A log-rank (Mantel‒Cox) test was performed to compare survival between groups (n = 5 mice). NS not significant.

Fig. 6. Therapeutic efficiency of intravesical instillation of AAV in orthotopic xenograft bladder tumor model mice.

a Schematic showing humanized orthotopic murine models of bladder cancer treated with intravesical instillation. Created in BioRender. Zhan, H. (2025) https://BioRender.com/oe2i6ny. b In vivo tumor bioluminescence images of each mouse treated with AAV-blank (n = 5 mice), BCG (n = 5 mice), mitomycin C (n = 5 mice), cMSC^SIPC/CtC^S (n = 5 mice), or gemcitabine + cisplatin (n = 6) from Day 0 (0 d) to Day 24 (24 d). c Tumor burdens of the mice in each group (n = 5, 5, 5, 5, and 6 mice per group). The data were presented as the mean ± SD, and significance was determined via two-tailed one-way ANOVA. d Representative images of tumors from each group described above. e Kaplan‒Meier survival curves of various groups. A log-rank (Mantel‒Cox) test was performed to compare survival between groups (n = 5, 5, 5, 5, and 6 mice per group). f Weights of the mice in each group (n = 5, 5, 5, 5, and 6 mice per group). The data were presented as the mean ± SD, and significance was determined via two-tailed Student’s t-tests. NS not significant. g–i The levels of serum IFN-γ (g), TNF-α (h), and IL-2 (i) in each group (n = 5, 5, 5, 5, and 5 mice, respectively) were determined by ELISA on Days 0, 7, 14, and 21. The data were presented as the mean ± SD, and significance was determined via two-tailed Student’s t-tests. NS not significant.