Search Agents face a severe reliability crisis during reinforcement learning (RL) fine-tuning. Heuristic Top-K retrieval often causes critical evidence loss or noise inclusion, while over-confidence induced by progressive RL leads to hallucinated answers and redundant searches. To build highly reliable agents, we introduce Conformal Prediction (CP) and propose Conformalized Agentic Search (CAS). This framework establishes reliability guarantees on both the retrieval and training sides: on the retrieval side, an Adaptive Prediction Set (APS), a specific CP realization, translates statistical coverage into dynamic document truncation to construct prediction sets that are adaptive in size; on the training side, Adaptive Conformal Inference (ACI), a dynamic CP algorithm, dynamically constructs prediction sets with controllable coverage to quantify answer confidence, which is then used to penalize low-confidence trajectories within the Group Relative Policy Optimization (GRPO) objective, ensuring the model learns only from reliable ones. Experiments across single-hop and multi-hop QA datasets demonstrate that our framework significantly improves reasoning accuracy while drastically reducing redundant tool invocations, establishing a highly reliable and efficient agent paradigm. Our code is available at https://github.com/S1llyBird/CAS.

Efficiently synthesizing novel views from sparse inputs while maintaining accuracy remains a critical challenge in 3D reconstruction. While advanced techniques like radiance fields and 3D Gaussian Splatting achieve rendering quality and impressive efficiency with dense view inputs, they suffer from significant geometric reconstruction errors when applied to sparse input views. Moreover, although recent methods leveraging monocular depth estimation to enhance geometric learning, their dependence on single-view estimated depth often leads to view inconsistency issues across different viewpoints. Consequently, this reliance on absolute depth can introduce inaccuracies in geometric information, ultimately compromising the quality of scene reconstruction with Gaussian splats. In this paper, we present RDG-GS, a novel sparse-view 3D rendering framework with Relative Depth Guidance based on 3D Gaussian Splatting. The core innovation lies in utilizing relative depth guidance to refine the Gaussian field, steering it towards view-consistent spatial geometric representations, thereby enabling the reconstruction of accurate geometric structures and capturing intricate textures. First, we devise refined depth priors to rectify the coarse estimated depth and insert global and fine-grained scene information into regular Gaussians. Building on this, to address spatial geometric inaccuracies from absolute depth, we propose relative depth guidance by optimizing the similarity between spatially correlated patches of depth and images. Additionally, we also directly deal with the sparse areas challenging to converge by the adaptive sampling for quick densification. Across extensive experiments on Mip-NeRF360, LLFF, DTU, and Blender, RDG-GS demonstrates state-of-the-art rendering quality and efficiency, making a significant advancement for real-world applications.

Cadmium (Cd) pollution poses a major threat to food safety. Sensitive detection of Cd2+ is of great significance for the life health. Herein, an aggregation-induced electrochemiluminescence (AIECL) sensor based on polymer dots (Pdots) is designed for Cd2+ detection. Through linkage of Cd2+ aptamer containing black hole quencher (BHQ) with Pdots, a signal "off" probe was achieved. In the presence of Cd2+, the BHQ moiety is released from Pdots and the ECL signal of Pdots is recovery. Such sensor exhibits excellent detection capability for Cd2+, with low detection limit of 0.006 ppb. More importantly, the sensor is applied for the quantitative analysis of Cd2+ in Ganoderma lucidum, and the results coincide well with the golden standard of ICP-MS. This study presents a facile and reliable methodology for the sensitive detection of cadmium, and demonstrates its potentials applying in food safety and environmental monitoring.

Medical Visual Question Answering (Medical-VQA) aims to to answer clinical questions regarding radiology images, assisting doctors with decision-making options. Nevertheless, current Medical-VQA models learn cross-modal representations through residing vision and texture encoders in dual separate spaces, which lead to indirect semantic alignment. In this paper, we propose UnICLAM, a Unified and Interpretable Medical-VQA model through Contrastive Representation Learning with Adversarial Masking. Specifically, to learn an aligned image-text representation, we first establish a unified dual-stream pre-training structure with the gradually soft-parameter sharing strategy. Technically, the proposed strategy learns a constraint for the vision and texture encoders to be close in a same space, which is gradually loosened as the higher number of layers. Moreover, for grasping the unified semantic representation, we extend the adversarial masking data augmentation to the contrastive representation learning of vision and text in a unified manner. Concretely, while the encoder training minimizes the distance between original and masking samples, the adversarial masking module keeps adversarial learning to conversely maximize the distance. Furthermore, we also intuitively take a further exploration to the unified adversarial masking augmentation model, which improves the potential ante-hoc interpretability with remarkable performance and efficiency. Experimental results on VQA-RAD and SLAKE public benchmarks demonstrate that UnICLAM outperforms existing 11 state-of-the-art Medical-VQA models. More importantly, we make an additional discussion about the performance of UnICLAM in diagnosing heart failure, verifying that UnICLAM exhibits superior few-shot adaption performance in practical disease diagnosis.

Ulcerative colitis, a type of inflammatory bowel disease, primarily impacts the colon's mucous membrane, leading to distressing symptoms. Both healthcare and socioeconomic aspects are significantly impacted by this condition. In this study, we stabilized proanthocyanidins (PAC) with silk sericin (SS) to create SS/PAC nanoparticles. These nanoparticles were then loaded into a thermosensitive in situ hydrogel made with Poloxamer. The effectiveness and safety of the resulting SS/PAC/P thermosensitive hydrogel were evaluated using a dextran sulfate sodium (DSS)-induced model of ulcerative colitis, through rectal administration of the hydrogel. SS/PAC/P demonstrated excellent compatibility and hemocompatibility. It exhibited strong antioxidant and anti-inflammatory properties, effectively relieving ulcerative colitis by counteracting its adverse effects. It alleviated abnormal disease activity index (DAI) scores, improved colon conditions, enhanced histological characteristics, and regulated microbial homeostasis. The superior effectiveness of SS/PAC/P hydrogel in vivo can be attributed to its prolonged residence time at the targeted site of administration, facilitating the continuous release of SS/PAC from the hydrogel. SS/PAC/P has shown potential as a promising therapeutic intervention for managing ulcerative colitis.

Recent large language models have become popular for achieving state-of-the-art performance in various natural language processing tasks, especially in zero-shot applications where tine-tuning is not required. However, these models underperform in text classification compared to tine-tuned models, due to limitations in reasoning ability and token constraints in in-context learning. Although extensive research explores large language models for text classification, few studies address multi-class classification with these models. This study introduces a new multi-class classification framework using a hierarchical label filtering strategy to manage long prompts in text classification. The influence of label sequence on classification accuracy is further investigated, demonstrating that optimized label arrangements significantly boost performance. Additionally, the study compares the performance of human-defined and model-generated labels in text classification, and analyze performance disparities across models for the same classification task.

Objectives: Inflammatory bowel disease (IBD), sepsis, and intestinal tumors are major health threats. This study aimed to explore the regulatory role of CD177+ neutrophils in the BMP signaling pathway and its impact on the onset, progression, and treatment of these diseases. Methods: Gene expression data from the Gene Expression Omnibus (GEO) database for IBD and sepsis were retrieved. Bioinformatics methods like background correction, normalization, and differential expression analysis were used. Weighted gene co-expression network analysis (WGCNA), gene functional enrichment analysis, pan-cancer analysis, single-cell analysis, and in vitro experiments including Caco-2 cell culture, cell proliferation assay (CCK-8), flow cytometry apoptosis analysis, quantitative real-time PCR (qRT-PCR), and plate colony formation assay were performed. Results: Key genes associated with IBD and sepsis, such as BMP2, BMP4, BMP6, BMP8A, and CD177, were identified. WGCNA in sepsis found two significant modules related to key clinical outcomes. Core gene screening revealed seven shared genes between IBD and sepsis, and enrichment analysis showed involvement in important biological processes and pathways. Pan-cancer analysis showed diverse gene expression patterns and correlations with immune dynamics. Single-cell transcriptomics provided insights into the tumor microenvironment. In vitro experiments demonstrated that CD177 knockdown affected BMP signaling pathway-related gene expression, ROS production, apoptosis, and cell proliferation. Conclusion: CD177+ neutrophils play a crucial role in regulating the BMP signaling pathway in IBD, sepsis, and intestinal tumors. These findings offer potential therapeutic targets, but further clinical validation is required to translate them into effective treatment strategies.

Medical vision-language pre-training (Med-VLP) models have recently accelerated the fast-growing medical diagnostics application. However, most Med-VLP models learn task-specific representations independently from scratch, thereby leading to great inflexibility when they work across multiple fine-tuning tasks. In this work, we propose UniDCP, a Unified medical vision-language model with Dynamic Cross-modal learnable Prompts, which can be plastically applied to multiple medical vision-language tasks within a unified model. Specifically, we explicitly construct a unified framework to harmonize diverse inputs from multiple pre-training tasks by leveraging cross-modal prompts for unification, which accordingly can accommodate heterogeneous medical fine-tuning tasks within a same model. Furthermore, we conceive a dynamic cross-modal prompt optimizing strategy that optimizes the prompts within the shareable space for implicitly processing the shareable clinic knowledge. UniDCP is the first Med-VLP model capable of performing all 8 medical uni-modal and cross-modal tasks over 14 corresponding datasets, consistently yielding superior results over diverse state-of-the-art methods.

Clinical evidence supports the notion that T cell exhaustion and terminal differentiation pose challenges to the persistence and effectiveness of chimeric antigen receptor-T (CAR-T) cells. MEK1/2 inhibitors (MEKIs), widely used in cancer treatment due to their ability to inhibit aberrant MAPK signaling, have shown potential synergistic effects when combined with immunotherapy. However, the impact and mechanisms of MEKIs on CAR-T cells remain uncertain and controversial. To address this, we conducted a comprehensive investigation to determine whether MEKIs enhance or impair the efficacy of CAR-T cells. Our findings revealed that MEKIs attenuated CAR-T cell exhaustion and terminal differentiation induced by tonic signaling and antigen stimulation, thereby improving CAR-T cell efficacy against hematological and solid tumors. Remarkably, these effects were independent of the specific scFvs and costimulatory domains utilized in CARs. Mechanistically, analysis of bulk and single-cell transcriptional profiles demonstrates that the effect of MEK inhibition was related to diminish anabolic metabolism and downregulation of c-Fos and JunB. Additionally, the overexpression of c-Fos or JunB in CAR-T cells counteracted the effects of MEK inhibition. Furthermore, our Cut-and-Tag assay revealed that MEK inhibition downregulated the JunB-driven gene profiles associated with exhaustion, differentiation, anergy, glycolysis, and apoptosis. In summary, our research unveil the critical role of the MAPK-c-Fos-JunB axis in driving CAR-T cell exhaustion and terminal differentiation. These mechanistic insights significantly broaden the potential application of MEKIs to enhance the effectiveness of CAR-T therapy.

Smartphones equipped with highly integrated sensors are increasingly being recognized as powerful tools for rapid on-site testing. Here, we propose a low-cost, portable, and highly multiplexed smartphone-based spectrometer capable of collecting three types of spectra—transmission, reflection, and fluorescence—by simply replacing the optical fiber attached to the housing. Spectral analysis is performed directly on the smartphone using a custom-developed app. Furthermore, we introduce a high signal-to-noise ratio (SNR) caffeine detection scheme that leverages aspirin and salicylic acid as fluorescent probes, allowing for the rapid and straightforward detection of caffeine in various samples. The fluorescence quenching of the probes was found to be linearly related to the caffeine concentration (0–200 μM), and the recoveries of the commercially available caffeine-containing samples were in the range of 98.0333–105.6000%, with a limit of detection (LOD) of 2.58 μM. The reliability and stability of the on-site assay using the smartphone spectrometer were verified. More importantly, this spectrometer demonstrates great potential as a versatile device for use outside of laboratory settings by enabling different operating modes tailored to various scenarios.

Medical generative models, acknowledged for their high-quality sample generation ability, have accelerated the fast growth of medical applications. However, recent works concentrate on separate medical generation models for distinct medical tasks and are restricted to inadequate medical multi-modal knowledge, constraining medical comprehensive diagnosis. In this paper, we propose MedM2G, a Medical Multi-Modal Generative framework, with the key innovation to align, extract, and generate medical multi-modal within a unified model. Extending beyond single or two medical modalities, we efficiently align medical multi-modal through the central alignment approach in the unified space. Significantly, our framework extracts valuable clinical knowledge by preserving the medical visual invariant of each imaging modal, thereby enhancing specific medical information for multi-modal generation. By conditioning the adaptive cross-guided parameters into the multi-flow diffusion framework, our model promotes flexible interactions among medical multi-modal for generation. MedM2G is the first medical generative model that unifies medical generation tasks of text-to-image, image-to-text, and unified generation of medical modalities (CT, MRI, X-ray). It performs 5 medical generation tasks across 10 datasets, consistently outperforming various state-of-the-art works.

Background Acute myeloid leukemia (AML) with biallelic ( CEBPA bi ) as well as single mutations located in the bZIP region is associated with a favorable prognosis, but the underlying mechanisms are still unclear. Here, we propose that two isoforms of C/EBPα regulate DNA damage-inducible transcript 3 (DDIT3) transcription in AML cells corporately, leading to altered susceptibility to endoplasmic reticulum (ER) stress and related drugs. Methods Human AML cell lines and murine myeloid precursor cell line 32Dcl3 cells were infected with recombinant lentiviruses to knock down CEBPA expression or over-express the two isoforms of C/EBPα. Quantitative real-time PCR and western immunoblotting were employed to determine gene expression levels. Cell apoptosis rates were assessed by flow cytometry. CFU assays were utilized to evaluate the differentiation potential of 32Dcl3 cells. Luciferase reporter analysis, ChIP-seq and ChIP-qPCR were used to validate the transcriptional regulatory ability and affinity of each C/EBPα isoform to specific sites at DDIT3 promoter. Finally, an AML xenograft model was generated to evaluate the in vivo therapeutic effect of agents. Results We found a negative correlation between CEBPA expression and DDIT3 levels in AML cells. After knockdown of CEBPA , DDIT3 expression was upregulated, resulting in increased apoptotic rate of AML cells induced by ER stress. Cebpa knockdown in mouse 32Dcl3 cells also led to impaired cell viability due to upregulation of Ddit3, thereby preventing leukemogenesis since their differentiation was blocked. Then we discovered that the two isoforms of C/EBPα regulate DDIT3 transcription in the opposite way. C/EBPα-p30 upregulated DDIT3 transcription when C/EBPα-p42 downregulated it instead. Both isoforms directly bound to the promoter region of DDIT3. However, C/EBPα-p30 has a unique binding site with stronger affinity than C/EBPα-p42. These findings indicated that balance of two isoforms of C/EBPα maintains protein homeostasis and surveil leukemia, and at least partially explained why AML cells with disrupted C/EBPα-p42 and/or overexpressed C/EBPα-p30 exhibit better response to chemotherapy stress. Additionally, we found that a low C/EBPα p42/p30 ratio induces resistance in AML cells to the BCL2 inhibitor venetoclax since BCL2 is a major target of DDIT3. This resistance can be overcome by combining ER stress inducers, such as tunicamycin and sorafenib in vitro and in vivo. Conclusion Our results indicate that AML patients with a low C/EBPα p42/p30 ratio (e.g., CEBPA bi ) may not benefit from monotherapy with BCL2 inhibitors. However, this issue can be resolved by combining ER stress inducers.

Myeloid-derived suppressor cells (MDSCs), a population of myeloid lineage cells with immunosuppressive capacity, can mitigate acute graft-versus-host disease (aGVHD) after allogeneic hematopoietic stem cell transplantation (allo-HSCT). We previously found that the immunosuppressive function of polymorphonuclear population (PMN-MDSCs) was impaired in aGVHD milieu. The aim of this study was to explore the intrinsic mechanism regulating the fate and function of donor-derived PMN-MDSCs during allo-HSCT. We firstly found that mitochondrial permeability transition pore (MPTP) opened in the PMN-MDSCs in response to the intense inflammatory environment of aGVHD, which induced mitochondrial damage, oxidative stress, and apoptosis of PMN-MDSCs. Inhibiting MPTP opening by a traditional immunosuppressant, cyclosporine A (CsA), could restore the immunosuppressive function and viability of PMN-MDSCs in vitro and in vivo, which reveals a new mechanism of CsA application.

A large variety of real engineering systems operate with multiple performance measures that are multistate in nature. These systems are usually modeled as multiperformance multistate systems (MPMSSs). However, existing MPMSS models fail to consider an important aspect, i.e., the performance conversion process. For example, in a combined heat and power (CHP) generating unit, apart from the output heat and electricity, decision-makers are also interested in the unit's capacity to convert gas into electricity and heat. The latter is related to the performance conversion process. This article proposes a framework for the reliability evaluation of performance conversion-based MPMSS. In the proposed MPMSS model, the couplings among different types of performances inside the components are quantified into the multistate performance conversion matrix. The performance conversion structure functions are proposed to derive system performance conversion capability based on the conversion capabilities of the components. Two reliability evaluation methods considering the steady-state performance conversion process and the continuous-time performance conversion process are proposed, respectively. Numerical examples are given to demonstrate the developed methods.

The gut microbiota and metabolites play pivotal roles in the pathobiology of various diseases. Here, we describe a protocol to profile the gut microbiome and meta-metabolome of a mouse disease model for acute graft-versus-host dis-ease. We describe steps for fecal sample collection and processing for 16S sequencing and UPLC-MS. Finally, we detail the steps for data analysis and exhibit multi-omic associations to correlate with pathology. For complete details on the use and execution of this protocol, please refer to Li et al. (2020).

Although widely applied in treating hematopoietic malignancies, transplantation of hematopoietic stem/progenitor cells (HSPCs) is impeded by HSPC shortage. Whether circulating HSPCs (cHSPCs) in steady-state blood could be used as an alternative source remains largely elusive. Here we develop a three-dimensional culture system (3DCS) including arginine, glycine, aspartate, and a series of factors. Fourteen-day culture of peripheral blood mononuclear cells (PBMNCs) in 3DCS led to 125- and 70-fold increase of the frequency and number of CD34 + cells. Further, 3DCS-expanded cHSPCs exhibited the similar reconstitution rate compared to CD34 + HSPCs in bone marrow. Mechanistically, 3DCS fabricated an immunomodulatory niche, secreting cytokines as TNF to support cHSPC survival and proliferation. Finally, 3DCS could also promote the expansion of cHSPCs in patients who failed in HSPC mobilization. Our 3DCS successfully expands rare cHSPCs, providing an alternative source for the HSPC therapy, particularly for the patients/donors who have failed in HSPC mobilization.

Introduction: The effects of MEK inhibitors (MEKIs) in CAR-T cells are poorly understood and remain controversial. Some groups showed that MEKIs could impair CAR-T cells' function in vitro. However, another group reported that MEKIs could combine with GD2 CAR-T cells to provide additional efficacy. Yet, the mechanism for this combined effect is unclear. Considering that the CAR signaling pathway is similar to TCR, both involve the activation of MAPK signaling. We hypothesized that MEKIs might mitigate CAR-T cells' exhaustion and terminal differentiation by lessening redundant CAR signaling. The study aims to systematically evaluate the role and mechanism of MEK inhibitors on CAR-T cells. Methods and Results: To explore whether MEKIs could mitigate the unbeneficial impact of antigen-independent CAR tonic signaling, we added 3 FDA-approved MEKIs: trametinib, cobimetinib, and binimetinib, respectively, to the culture medium of CD19.28z CAR-T cells for 9 days at the concentration approaching their clinically tolerable peak blood concentration. We found that all the 3 MEKIs could reduce CAR-T cells' terminal differentiation and restrain the expression of exhaustion and activation markers. Among them, trametinib was the most potent because it could achieve the comparable effect of cobimetinib and binimetinib at the lowest concentration. Thus, we chose trametinib for further research. We confirmed that compared to concentrations of 7.5nM and 30nM, 15nM was the optimal concentration of trametinib with mild inhibition on the proliferation of CAR-T cells and a potent effect on CAR-T cells' phenotype. Pre-treatment with trametinib didn't affect the in-vitro cytotoxicity of CD19.28z CAR-T cells. The above effects of trametinib were more significant as treatment time and the dose increased. Similar results could be obtained in analogous experiments with CD19.4-1BBz and GD2.28z CAR-T cells. Intriguingly, GD2.28z CAR-T cells cultured with trametinib had better proliferation and killing capacity because they were more exhausted than CD19.28z CAR-T cells. Using the Nalm-6-bearing leukemia xenograft model, we found that compared to DMSO pre-treated CD19.28z CAR-T cells, trametinib pre-treated CD19.28z CAR-T cells exerted more potent anti-leukemia activity, showed a less exhausted and differentiated state, proliferated better, and further extended mice survival (Fig.1). Similar results could be obtained in analogous in-vivo experiments with CD19.4-1BBz CAR-T cells. To evaluate whether trametinib could protect CAR-T cells from the exhaustion and terminal differentiation triggered by antigen stimulation, we cocultured CD19.28z CAR-T cells with Nalm-6 cells in a medium with or without trametinib. We demonstrated under antigen stimulation, trametinib could effectively inhibit CAR-T cells' activation, exhaustion, apoptosis, terminal differentiation, and phosphorylation of ERK and consequently promote the proliferation of total and CD8 CAR-T cells. In addition, after repetitive antigen stimulation, trametinib could rescue CAR-T cells' functional exhaustion and improve CAR-T cells' in-vitro cytotoxicity. Mechanistically, Single-cell and bulk RNA-Seq revealed that the effect of MEK inhibition was associated with the downregulation of AP-1 and exhaustion-associated transcription factors (TFs) and the upregulation of memory-associated TFs (Fig.2). GSEA revealed the upregulation of naive/memory-associated genes and downregulation of genes involved in T cell activation/effector/exhaustion, AP-1 pathway and apoptosis in the trametinib-treated group. Additionally, single-cell transcriptional profiling demonstrated enrichment of memory and Ki67+cycling CAR-T clusters in the trametinib-treated group. Among the AP-1 TFs downregulated by trametinib, c-Fos and JunB are the direct downstream targets of the canonical MAPK signaling pathway, and both of them were reported to implicate T cell exhaustion. Thus we speculated both downregulation of c-Fos and JunB may contribute to the role of MEKIs. Consistent with our hypothesis, overexpression of c-Fos or JunB in CAR-T cells could partially, if not all, abrogated the effects of MEK inhibition. Significance: Our research provides a strategy to optimize CAR-T antitumor efficacy by MEK inhibition and shed light on the role of c-Fos and JunB in driving CAR-T cell exhaustion and terminal differentiation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Myeloid-derived suppressor cells (MDSCs) represent a population of heterogeneous myeloid cells, which are characterized by their remarkable ability to suppress T cells and natural killer cells. MDSCs have been proven to play a positive role in protecting acute graft-versus-host disease (aGVHD). Here, we aimed to describe the mechanism behind how mTOR signaling regulates MDSCs’ generation and explore its prophylactic and therapeutic potential in aGVHD. Reducing mTOR expression retains myeloid cells with immature characteristics and promotes polymorphonuclear MDSC (PMN-MDSC) immunosuppressive function through STAT3-C/EBPβ pathway. Prophylactic transfusion of mTORKO PMN-MDSCs could alleviate aGVHD while maintaining the graft-versus-leukemia (GVL) effect, which could downregulate the Th1/Th2 ratio, decrease serum proinflammatory cytokines, and increase the proportion of regulatory T cells (Tregs) in aGVHD models at the early stage after transplantation. Moreover, transfusion therapy could promote the reconstruction and function of donor-derived PMN-MDSCs. Not only the percentage and the absolute number of donor-derived PMN-MDSCs significantly increased but also the immunosuppressive ability was much more robust compared to other groups. Altogether, these findings indicated that mTOR is an intrinsic regulator for PMN-MDSCs’ differentiation and immunosuppressive function. Together, mTORKO PMN-MDSC transfusion can play a protective role in alleviating cytokine storm at the initial stage and promoting the quantitative and functional recoveries of donor-derived PMN-MDSCs in aGVHD.

Purpose Because of the dose-dependent increased risk of cardiovascular events we tried to lower the dose of ponatinib without reducing its efficacy in the treatment of BCR-ABL T315I-containing CML. Combination with hydroxychloroquine can enhance the efficiency of ponatinib and axitinib through autophagy inhibition in CML cell with T315I. Methods Cell viability, cell cycle, cellular senescence, formation of cell clones and apoptosis assay were taken to test the efficiency of medicine. Lentiviral vectors containing shRNA was used to block autophagy and to verify the mechanism of the medicine. Establish tumor models in nude mice, and verify the experimental results in vivo. Results ponatinib and axitinib killed 32Dp210-T315I cells as well as inducing autophagy, which promoted their survival under pressure from TKIs. By inhibiting autophagy, HCQ enhanced the killing effect of ponatinib and axitinib on 32Dp210-T315I cells. In vivo HCQ also enhanced the killing effect of axitinib on 32Dp210-T315I cells. Conclusion HCQ combined with ponatinib may be a new strategy for treating CML and ALL that harbor the T315I mutation. Thus, this combination may make it possible to reduce the dose of ponatinib and reduce its side effects without compromising efficacy.