Predicting molecular bioactivity in low-data regimes remains a central challenge in computational drug discovery, where labeled compounds for specialized tasks are scarce while related datasets are abundant. Here, we propose FerrGAT, a graph attention network framework that addresses this challenge through domain-relevant multi-task pre-training and dual-channel molecular representation learning. FerrGAT first pre-trains a shared GAT encoder on three mechanistically related tasks—antioxidant activity (GST inhibition, 245 compounds from ChEMBL target CHEMBL2095173; 83 active/162 inactive), kinase inhibition (3000 compounds spanning AXL, EGFR and VEGFR2 from ChEMBL; 2240 active/760 inactive), and cellular toxicity (7265 compounds from the Tox21 NR-AhR endpoint; 309 active/6956 inactive)—then transfers the learned representations to a target task via differential learning rate fine-tuning. The architecture fuses atom-level graph features from multi-head attention message passing with global physicochemical descriptors through a learned projection and provides built-in interpretability via attention weight visualization at the atomic level. We evaluated FerrGAT on ferroptosis inhibitor prediction as a representative low-data molecular classification task (1052 compounds from ChEMBL targets GPX4 and HMOX1 combined with 63 literature- and FerrDb-curated ferroptosis modulators; 409 active/643 inactive). In 5-fold cross-validation, FerrGAT achieved an AUC of 0.906, outperforming Morgan fingerprint baselines, including Random Forest (0.877), XGBoost (0.871), and an SVM (0.873). Ablation studies confirmed that domain-relevant pre-training improved AUC by 2.5% over training from scratch, while pre-training on unrelated tasks degraded performance, highlighting the importance of task-domain alignment. Applied to virtual screening of 30 FDA-approved tyrosine kinase inhibitors, the model identified Bemcentinib (AXL inhibitor, score = 0.918) as a top candidate, validated by AutoDock Vina molecular docking (−8.51 kcal/mol) and independent experimental evidence, including lipid peroxidation assays, Western blot, and cellular thermal shift analysis. These results demonstrate that domain-aware transfer learning with graph attention networks provides an effective and interpretable framework for molecular property prediction in data-limited scenarios.

Current lysosome-targeting chimeras (LYTACs) and antibody-drug conjugates (ADCs) face inherent limitations, including reliance on monospecific binders and target-dependent internalization efficiency. Here, we report a versatile therapeutic platform that overcomes these constraints by engineering a novel class of multivalent, multispecific binding proteins, termed Multibodies . Using the fungal immunomodulatory protein LZ-8 as a scaffold, we developed a lysosome-targeting chimera-drug conjugates (LYTAC-DCs) system capable of mediating the degradation of multiple membrane proteins while enabling site-specific drug release. While wild-type LZ-8 demonstrated potent lysosomal targeting, it induced significant off-target toxicity and lysosomal dysfunction. Through rational protein engineering, we identified a key residue (Y84) responsible for promiscuous receptor binding and generated an optimized variant, LZ-8-2.3 (Y84A), which eliminated toxicity while preserving efficient lysosomal trafficking. The resulting LYTAC-DC platform mediated high-fidelity degradation of EGFR, PD-L1, and HER2 across diverse cancer models, concurrently delivering cytotoxic (e.g., MMAE) or immunogenic (e.g., doxorubicin) payloads. Efficacy was validated in patient-derived organoids and murine xenografts, including against osimertinib-resistant lung cancer. Furthermore, AI-assisted directed evolution enabled the development of non-chimeric lysosome-targeting drug conjugates (LYTA-DCs), highlighting the modularity and engineerability of the Multibody scaffold. Our work establishes a unified and programmable strategy for targeted protein degradation and drug delivery, significantly expanding the therapeutic landscape beyond conventional LYTAC and ADC technologies.

Cytochrome C (Cyt C) is a central mediator of intrinsic apoptosis, whereas its heme-free precursor, apocytochrome C (APO-Cyt C), competitively inhibits this process. Sodium aescinate (SA), a natural triterpene saponin, is known to facilitate the endosomal escape of protein drugs. We initially aimed to investigate the effects of enhancing endosomal escape efficiency on protein activity through the combination of SA and Cyt C. However, this study investigates an cytotoxicity observed when APO-Cyt C is combined with SA and aims to elucidate the underlying molecular mechanism. Contrary to its established anti-apoptotic function, APO-Cyt C, when co-administered with a non-toxic concentration of SA, induced potent, caspase-dependent mitochondrial apoptosis in cancer cells. This pro-apoptotic switch was not primarily triggered by BCL-2 family protein modulation, ROS generation, or calcium overload. Instead, the primary mechanism is the induction of excessive and lethal autophagy. SA was found to induce lysosomal membrane damage, evidenced by Galectin-9 recruitment, which initiates lysophagy. The addition of APO-Cyt C significantly amplified the autophagic flux, leading to decreased p62 levels and enhanced LC3-II turnover. Mechanistically, this synergy is driven by a dual impact on the AKT-mTOR-TFEB pathway: APO-Cyt C treatment decreased mTOR phosphorylation, while the combination promoted the nuclear translocation of the autophagy regulator, TFEB. Inhibition of autophagic flux using Bafilomycin A1 or Tetrandrine rescued cells from apoptosis, confirming that excessive autophagy is the direct cause of cell death. This study reveals a novel therapeutic strategy wherein an anti-apoptotic protein is converted into a potent pro-apoptotic agent. The combination of APO-Cyt C and SA triggers apoptosis by overwhelming the cell with excessive autophagic flux, driven by synergistic inhibition of the mTOR-TFEB axis. These findings highlight the therapeutic potential of modulating autophagy and suggest that combining mTOR inhibitors with lysosome-targeting agents like SA could be an effective anti-cancer strategy.

Extracellular targeted protein degradation (eTPD) systems typically utilize lysosome-targeting receptors (LTRs) to mediate internalization and lysosomal degradation of extracellular and membrane proteins. While multiple LTRs have been discovered, there remains a compelling need to seek for new LTRs, particularly those with clear clinical relevance, to expand the therapeutic potential of eTPD. Here we report trophoblast cell surface antigen-2 (TROP2), a clinically validated tumor-associated antigen, as a promising tumor-selective LTR. We engineer TROP2-targeting chimeras (TRTACs) by genetically fusing a TROP2-binding nanobody to nanobodies against specific target proteins. We show that TRTACs can induce tumor cell-selective degradation of diverse membrane proteins, including epithelial growth factor receptor (EGFR), human epithelial growth factor receptor 2 (HER2), and programmed death-ligand 1 (PD-L1). The EGFR-targeted TRTAC significantly inhibits tumor cell proliferation and shows potent antitumor activity in vivo. We further design TRTAC-drug conjugates (TRTAC-DCs) by attaching cytotoxic payloads to TRTACs, enabling targeted protein degradation together with enhanced drug delivery. TRTAC-DCs show significantly enhanced activity against HER2- and EGFR-positive tumors both in vitro and in vivo, with minimal toxicity observed in normal tissues. These findings establish TROP2 as a robust LTR and provide a versatile eTPD platform with profound translational potential for tumor treatment.

The nonspecific toxicity of traditional chemotherapeutic agents and the intracellular delivery barriers of protein-based drugs have limited their clinical applications. Efficient targeted delivery properties of toxin proteins themselves provide important insights into the design of novel drug delivery systems. Inspired by the natural targeting properties of the plant-derived type I ribosome-inactivating protein (RIPs) MAP30, we engineered a detoxified carrier, MAP30ER, through site-directed mutagenesis of key enzymatic residues (E158A/R161A). This variant retained its cell-binding capability while exhibiting significantly reduced cytotoxicity. ELISA and molecular docking identified EGFR as the primary functional receptor of MAP30, and flow cytometry confirmed a 3-fold higher binding affinity of MAP30 to EGFR-high tumor cells (A431) compared to EGFR-low cells (HeLa). To address endosomal entrapment, we screened triterpenoid saponins and found that Saikosaponin A (SSA) enhanced the endosomal escape efficiency of MAP30ER by inducing vesicle membrane disruption. When applied to deliver Apoptin (a apoptosis-inducing protein), the MAP30ER-SSA system elevated tumor cell killing efficacy from 65 % to 90 % upon SSA co-treatment, and Western blot confirmed that it induced apoptosis through activation of caspase-3/9 and cleavage of PARP1. Furthermore, we extended antibody-drug conjugate (ADC) technology to the MAP30ER platform, constructing a MAP30ER-MMAE conjugate that achieved targeted delivery to A431 cells with negligible toxicity to normal cells. This study establishes a novel paradigm for engineering plant-derived RIPs and provides a multifunctional delivery platform with targeting specificity, safety, and versatility, offering innovative solutions to overcome drug delivery challenges in cancer therapy.

Most tumors are resistant to programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) checkpoint inhibitors, which may be due to impaired antigen presentation resulting from the downregulation of major histocompatibility complex class I (MHC-I) expression on tumor cells. We observed that platycodin D (PD), polygalacin D, and platycodin D2, which are plant-derived triterpenoid saponins, significantly reduced PD-L1 levels. RNA sequencing and the PharmMapper database analysis identified liver X receptor beta (LXR-beta) as a potential PD target. Further studies showed that PD reduces PD-L1 levels by binding to LXR-beta and inhibiting LXR-beta activity. Coadministration of PD and nintedanib, known to upregulate MHC-I expression, enhanced tumor recognition and killing by T cells. This study provides new insights into PD applications and mechanisms.

The susceptibility of lysosomal membranes in tumor cells to cationic amphiphilic drugs (CADs) enables CADs to induce lysosomal membrane permeabilization (LMP) and trigger lysosome-dependent cell death (LDCD), suggesting a potential antitumor therapeutic approach. However, the existence of intrinsic lysosomal damage response mechanisms limits the display of the pharmacological activity of CADs. In this study, we report that low concentrations of QS-21, a saponin with cationic amphiphilicity extracted from Quillaja Saponaria tree, can induce LMP but has nontoxicity to tumor cells. QS-21 and MAP30, a type I ribosome-inactivating protein, synergistically induce apoptosis in tumor cells at low concentrations of both. Mechanistically, QS-21-induced LMP helps MAP30 escape from endosomes or lysosomes and subsequently enter the endoplasmic reticulum, where MAP30 downregulates the expression of autophagy-associated LC3 proteins, thereby inhibiting lysophagy. The inhibition of lysophagy results in the impaired clearance of damaged lysosomes, leading to the leakage of massive lysosomal contents such as cathepsins into the cytoplasm, ultimately triggering LDCD. In summary, our study showed that coadministration of QS-21 and MAP30 amplified the lysosomal disruption and can be a new synergistic LDCD-based antitumor therapy.

Cell‐penetrating peptides (CPPs) with better biomolecule delivery properties will expand their clinical applications. Using the MLCPP2.0 machine algorithm, we screened multiple candidate sequences with potential cellular uptake ability from the nuclear localization signal/nuclear export signal database and verified them through cell‐penetrating fluorescent tracing experiments. A peptide (NCR) derived from the Rev protein of the caprine arthritis‐encephalitis virus exhibited efficient cell‐penetrating activity, delivering over four times more EGFP than the classical CPP TAT, allowing it to accumulate in lysosomes. Structural and property analysis revealed that a high hydrophobic moment and an appropriate hydrophobic region contribute to the high delivery activity of NCR. Trastuzumab emtansine (T‐DM1), a HER2‐targeted antibody‐drug conjugate, could improve its anti‐tumor activity by enhancing targeted delivery efficiency and increasing lysosomal drug delivery. This study designed a new NCR vector to non‐covalently bind T‐DM1 by fusing domain Z, which can specifically bind to the Fc region of immunoglobulin G and effectively deliver T‐DM1 to lysosomes. MTT results showed that the domain Z‐NCR vector significantly enhanced the cytotoxicity of T‐DM1 against HER2‐positive tumor cells while maintaining drug specificity. Our results make a useful attempt to explore the potential application of CPP as a lysosome‐targeted delivery tool.

To address the challenges posed by low immunogenicity and immune checkpoints during cancer treatment, we propose an alternative strategy that combines immunogenic cell death (ICD) effects with CD47/SIRPα blockade to reactivate phagocytosis of tumor cells by macrophages with polysaccharide-based drug delivery. In this study, the EGFR inhibitor gefitinib was identified as a novel CD47 modulator, which promoted the translocation of CD47 molecules from the cell membrane to endosomes through the EGFR-Rab5 pathway, leading to reduced cell surface CD47 levels and limiting interaction with SIRPα. Based on this finding, we developed prophagocytic mixed nanodrugs to enhance macrophage phagocytosis by encapsulating ICD inducer doxorubicin and CD47 inhibitor gefitinib with immunostimulatory polysaccharides from Ganoderma lucidum. This approach downregulated cell surface CD47 expression to attenuate "don't-eat-me" signaling, while increasing doxorubicin accumulation in tumors by inhibiting drug-resistance proteins, leading to more exposure of calreticulin and amplifying the "eat-me" signaling. In vivo experiments demonstrated that this approach significantly suppressed intraperitoneal tumor dissemination, reversed doxorubicin-induced weight loss, and effectively induced macrophage polarization, dendritic cell maturation, and CD8+ T cell activation. These findings highlighted the significant potential of our macrophage-centered therapeutic strategy using polysaccharide-based nanocarriers and provided new perspectives for chemoimmunotherapy.

Plant-derived triterpenoid saponins have been shown to play a powerful role in enhancing the cytotoxic activity of protein therapeutics. However, the mechanism of how saponins are acting is not clearly understood. In this study, momordin Ic (MIC), a triterpenoid saponin derived from Kochia scoparia (L.) Schrad., specifically enhance the antiproliferative effect of recombinant MAP30 (a type I ribosome inactivating protein, RIP) in breast cancer cells. Subsequently, the possible mechanism of how MIC enhanced the cytotoxicity of MAP30 was analyzed in detail. We observed the level of intracellular labeled MAP30 using fluorescence microscopy and flow cytometry. And a reporter protein, GAL9, was used to monitor the role of MIC in promoting endosomal escape. We found endosomal escape does not play a role for the enhancer effect of MIC while the effect of MIC on MAP30 is cholesterol dependent and that ganglioside GM1, a lipid raft marker, can competitively inhibit cytotoxicity of MAP30 enhanced by MIC. Finally, we provided some insights into the correlation between the sugar side chain of MIC and its role in enhancing of RIP cytotoxicity and altering of drug cell tropism.

Recently, therapeutic cancer vaccines have emerged as promising candidates for cancer immunotherapy. Nevertheless, their efficacies are frequently impeded by challenges including inadequate antigen encapsulation, insufficient immune activation, and immunosuppressive tumor microenvironment. Herein, we report a three-in-one hydrogel assembled by nucleic acids (NAs) that can serve as a vaccine to in situ trigger strong immune response against cancer. Through site-specifically grafting the chemodrug, 7-ethyl-10-hydroxycamptothecin (also known as SN38), onto three component phosphorothioate (PS) DNA strands, a Y-shaped motif (Y-motif) with sticky ends is self-assembled, at one terminus of which an unmethylated cytosine-phosphate-guanine (CpG) segment is introduced as an immune agonist. Thereafter, programmed cell death ligand-1 (PD-L1) siRNA that performs as immune checkpoint inhibitor is designed as a crosslinker to assemble with the CpG- and SN38-containing Y-motif, resulting in the formation of final NA hydrogel vaccine. With three functional agents inside, the hydrogel can remarkably induce the immunogenic cell death to enhance the antigen presentation, promoting the dendritic cell maturation and effector T lymphocyte infiltration, as well as relieving the immunosuppressive tumor environment. When inoculated twice at tumor sites, the vaccine demonstrates a substantial antitumor effect in melanoma mouse model, proving its potential as a general platform for synergistic cancer immunotherapy.

Targeted therapy has attracted more and more attention in cancer treatment in recent years. However, due to the diversity of tumor types and the mutation of target sites on the tumor surface, some existing targets are no longer suitable for tumor therapy. In addition, the long-term administration of a single targeted drug can also lead to drug resistance and attenuate drug potency, so it is important to develop new targets for tumor therapy. The expression of Type III transforming growth factor β receptor (TGFBR3) is upregulated in colon, breast, and prostate cancer cells, and plays an important role in the occurrence and development of these cancers, so TGFBR3 may be developed as a novel target for tumor therapy, but so far there is no report on this research. In this study, the structure of bone morphogenetic protein 4 (BMP4), one of the ligands of TGFBR3 was analyzed through the docking analysis with TGFBR3 and sequence charge characteristic analysis, and a functional tumor-targeting penetrating peptide T3BP was identified. The results of fluorescent labeling experiments showed that T3BP could target and efficiently enter tumor cells with high expression of TGFBR3, especially A549 cells. When the expression of TGFBR3 on the surface of tumor cells (HeLa) was knocked down by RNA interference, the high delivery efficiency of T3BP was correspondingly reduced by 40%, indicating that the delivery was TGFBR3-dependent. Trichosanthin (TCS, a plant-derived ribosome inactivating protein) fused with T3BP can enhance the inhibitory activity of the fusion protein on A549 cells by more than 200 times that of TCS alone. These results indicated that T3BP, as a novel targeting peptide that can efficiently bind TGFBR3 and be used for targeted therapy of tumors with high expression of TGFBR3. This study enriches the supply of tumor-targeting peptides and provides a new potential application option for the treatment of tumors with high expression of TGFBR3.

Targeted delivery of antitumor drugs is particularly important in tumour treatment. Tumour-targeted peptide is a very effective drug carrier for tumour therapy. Here, we screened and characterised a highly efficient targeted peptide named IHP5, which was derived from insulin-like growth factor binding proteins (IGFBPs). IHP5 exhibited preferential binding to the tested tumour cell lines. The delivery efficiency of IHP5 was higher in various tested tumour cells than in normal cells, especially in the human cervical cancer cell line HeLa, which was 11.7-fold higher than in normal human embryonic kidney cells HEK293. Moreover, the penetration efficiency of IHP5 was 13 times higher than that of the classical cell penetrating peptide TAT in HeLa cells. Detail analysis revealed that IHP5 endocytosis was possibly correlated with acetylated heparan sulphate proteoglycans including phosphatidylinositol proteoglycan 3 (GPC3), phosphatidylinositol proteoglycan 5 (GPC5) and syndecan 2 (SDC2). Subsequently, the introduction of IHP5 enhanced the inhibitory effect of trichosanthin (TCS) on tumour cells, resulting in at least 19-fold increase in tumour cells without enhanced cytotoxicity in normal cells HEK293. These results suggested that IHP5, as a novel tumour cell-targeting penetrating peptide with the ability to target tumour cells, has great potential in drug delivery applications.

Bifunctional chimeric protein containing PP1-disrupting sequence induced calreticulin exposure and reversed immunosuppressive tumor microenvironment, thereby enhancing antitumor effect.

Cell-penetrating peptides (CPPs) have been regarded as potential drug carriers for cancer therapy. However, most well-studied CPPs fail to deliver exogenous drugs efficiently and selectively. In this study, a tumour-targeted CPP with high efficiency derived from heparin-binding domain (HBD) of Midkine (named HMD) was discovered. HMD exhibited higher delivery efficiency than classic CPPs (TAT and R9) and manifested selectivity in tumour cells. Normally, the positive charge is the key factor for the transmembrane activity of CPPs such as TAT and R9. Here, the length of alpha-helix inside CPP was found also important for in the recognition of heparan sulphate proteoglycans (HSPGs). Subsequently, the introduction of HMD enhanced the inhibitory effect of Momordica antiviral protein of 30 kDa (MAP30) on tumour cells, resulting in a 6.07-fold and 5.42-fold increase in HeLa cells and MGC80-3 cells respectively without enhanced cytotoxicity in normal cells. These results show that HMD possesses high efficiency and good tumour specificity and can be utilised as a promising agent for the tumour-targeted delivery of drug. This study is also a supplement to the existing theories about the biological activities of the alpha-helix in CPPs.

Biomacromolecules such as proteins and nucleic acids are very attractive due to their high efficiency and specificity as cancer therapeutics. In fact, the endocytosed macromolecules are often trapped in the endosomes and cannot exhibit pharmacological effects well. Many strategies have been used to address this bottleneck, and one promising approach is to exploit the endosomal escape-promoting effect of triterpenoid saponins to aid in the release of biomacromolecules. Here, Raddeanin A (RA, an oleanane-type triterpenoid saponin) was proved to significantly promote endosomal escape as it recruited Galectin-9, an endosomal escape event reporter. As expected, RA effectively enhanced the anti-tumor effect of MAP30 (a type I ribosome-inactivating protein derived from Momordica charantia). However, based on the results of fluorescent colocalization, RA did not significantly promote MAP30 release from endosomes, suggesting that RA enhances MAP30 activity not only by promoting endosomal escape. Furthermore, it was found that the inhibitors of micropinocytosis and caveolae could almost completely inhibit the cytotoxicity of MAP30 combined with RA without affecting the cytotoxicity of MAP30 alone, indicating that RA may regulate the endocytic pathway of MAP30. Meanwhile, the effect of RA is related to the intra vesicular pH and cholesterol content on cell membrane, and is also cell-type dependent. Therefore, RA enhanced the anti-tumor effect of MAP30 in multiple ways, not just by promoting endosomal escape. Our findings will help to further decipher the possible mechanisms by which triterpenoid saponins enhance drug activity, and provide a new perspective for improving the activity of endocytosed drugs.

Trichosanthin (TCS), as a type 1 ribosome-inactivating protein, has a very high cytoplasmic activity in vitro and can quickly kill cancer cells. However, it is easily filtered and cleared by the kidney, which results in the short half-life and severely limits its application. In this study, we constructed several recombinant proteins by fusing the albumin binding domain mutant ABD035(abbreviated as ABD) to the N-or C-terminus of TCS to endow the recombinant TCS fusion protein with a longer half-life property binding with endogenous human serum albumin (HSA) via ABD to effectively exert its anti-tumor activity in vivo. Pull down, Dynamic light scattering and ELISA assays all showed that TCS fused with two ABD sequences at the C-terminus of TCS, has stronger binding capacity to HSA in vitro than TCS with one ABD. In vivo studies in BALB/C mice were performed and the elimination half-life of TCS-ABD-ABD is about 15-fold longer compared to TCS and anti-tumor activity is about 30% higher than that of TCS alone in BALB/C mouse experiments. Moreover, we found that TCS with two ABDs in tandem have the highest soluble expression level, more than 5 times higher than that of TCS, and the yield of purified protein of TCS-ABD-ABD was as high as 68.9 mg/L culture solution, which was about 7-fold higher than that of TCS. Furthermore, MTT assay showed that the anti-tumor activity of TCS-ABD-ABD was significantly higher than TCS fused with only one ABD sequence, indicating that the repeated ABD sequences facilitated the biological activity of TCS. In this paper, the fusion of the albumin-binding domain in tandem with TCS can effectively improve its stability in vivo and also significantly increase its soluble expression, expanding the application of the albumin -binding domain in the high soluble expression and stability of protein drugs.

Ovalbumin (OVA) is a model antigen commonly incorporated in smartly designed nanoparticles for delivery into antigen-presenting cells (APC), aiming to investigate the immune activity and therapeutic efficacy of nanoparticles that contain immunoregulatory compounds. However, the immunoresponse observed in nano-immunotherapy may unexpectedly arise from endotoxin impurity of OVA in the nanoparticles. Literature review shows that most researchers did not notice the importance of endotoxin-free OVA when used in nano-immunotherapy studies. Concentration at as low as 5 μg/ml OVA from Sigma-Aldrich (contains 0.625 ng/ml endotoxin) was able to activate APC such as dendritic cells and macrophages. Here, we proposed that the endotoxin impurity in OVA or the finished nanoproducts should be determined by both Limulus Amebocyte Lysate (LAL) and cell-based assay, to ensure the endotoxin-free quality of the nanoparticles. The endotoxin in OVA can be removed by endotoxin removal column and phase separation methods and endotoxin-free OVA can be purchased. This perspective alerts the researchers of endotoxin impurity of OVA that may transfer into the finished nanoparticles and introduce an unfavorable immunoregulatory function with false-positive results. OVA with minimal endotoxin level should be used in nano-immunotherapy studies to accurately reflect the true effects of nanoparticles on the immune system. This article is categorized under: Toxicology and Regulatory Issues in Nanomedicine > Toxicology of Nanomaterials Nanotechnology Approaches to Biology > Nanoscale Systems in Biology.

Efficient endosomal escape after cellular uptake is a major challenge for the clinical application of therapeutic proteins. To overcome this obstacle, several strategies have been used to help protein drugs escape from endosomes without affecting the integrity of the cell membrane. Among them, some triterpenoid saponins with special structures were used to greatly enhance the anti-tumor therapeutic effect of protein toxins. Herein, we demonstrated that platycodin D (PD), polygalacin D (PGD) and platycodin D2 (PD2) from Platycodonis Radix significantly enhanced the ability of MHBP (a type I ribosome-inactivating protein toxin MAP30 fused with a cell-penetrating peptide HBP) to induce apoptosis in hepatoma cells. Based on the results of co-localization of endocytosed EGFP-HBP with a lysosomal probe and Galectin-9 vesicle membrane damage sensor, we demonstrated that PD, PGD and PD2 have the ability to promote endosomal escape of endocytic proteins without affecting the integrity of the plasma membrane. Meanwhile, we observed that cholesterol metabolism plays an important role in the activity of PD by RNA-seq analysis and KEGG pathway enrichment analysis, and confirm that PD, PGD and PD2 enhance the anti-tumor activity of MHBP by inducing the redistribution of free cholesterol and inhibiting the activity of cathepsin B and cathepsin D. Finally, we found that PD synergized with MHBP to induce caspase-dependent apoptosis through inhibiting Akt and ERK1/2 signaling pathways and activating JNK and p38 MAPK signaling pathways. This study provides new insights into the application of PD in cancer therapy and provides efficient and promising strategies for the cytosolic delivery of therapeutic proteins.

Targeted delivery of antitumor drugs is especially important for tumour therapy. Tumour targeting peptides have been shown to be very effective drug carriers for tumour therapy. Interleukin-4 receptor (IL-4R) is overexpressed on the surface of various human solid tumours. To obtain a better targeting peptide, we first designed a novel targeting peptide derived from interleukin-4 (IL-4), ILBP-b. ILBP-b contains the key high-affinity binding residue E9 of IL-4 to IL-4R. Compared with a reported targeting peptide ILBP-a (containing another key high affinity residue R88), ILBP-b was proved to be a better targeting peptide by the fluorescence experiments. Then, we further fused ILBP-b and ILBP-a to increase the multisite-binding ability of ILBP-b and got a better targeting peptide ILBP-ba. ILBP-ba showed a stronger preferential binding ability to IL-4R high-expressing cells than ILBP-a and ILBP-b. Competitive binding experiments demonstrated ILBP-ba specifically targets IL-4R. By fusing ILBP-ba with drug protein trichosanthin (TCS), in vitro drug carrying experiments showed that ILBP-ba could specifically enhance the killing effect of TCS on IL-4R high-expressing tumour cells (more than 10 folds). These results indicated that ILBP-ba has great potential for drug delivery applications targeting IL-4R and will be beneficial for the development of tumour therapeutic agents.