Explore the vast range of titles available.

Adenosine signaling is a crucial immunosuppressive pathway within the tumor microenvironment, making it a promising target for cancer therapy. In this study, it is demonstrated that Globo H ceramide (GHCer), the most prevalent tumor‐associated glycosphingolipid, influences the tumor microenvironment by activating adenosine signaling, which results in dual immunosuppressive effects on T cells. It is demonstrated that GHCer interacts with the adenosine receptor 2A (A2AR), triggering cyclic AMP (cAMP) and protein kinase A (PKA) signaling. This interaction leads to a reduction in the proliferation of CD4+ T cells while simultaneously promoting the differentiation of regulatory T cells (Tregs). Furthermore, GHCer enhances the suppressive capacity of Treg cells by upregulating inhibitory molecules such as Lymphocyte‐activation gene 3 (LAG3), cytotoxic T‐lymphocyte‐associated protein 4 (CTLA‐4), Programmed cell death 1 ligand 1 (PD‐L1), and it stimulates the secretion of the immunosuppressive cytokine Interleukin 35 (IL‐35). Additionally, GHCer‐induced Tregs express CD39 and CD73, which further enhances adenosine production and creates a positive feedback loop in the adenosinergic pathway and A2AR signaling. Mechanistically, it is found that GHCer forms a complex with TRAX (translin‐associated factor‐X) and the C‐terminus of A2AR, which facilitates the activation of A2AR and promotes an immunosuppressive tumor microenvironment. This study identifies Globo H ceramide (GHCer) as a novel immune checkpoint molecule that suppresses the activation of conventional T lymphocytes while promoting the differentiation and function of regulatory T cells. The immunosuppressive effects of GHCer are mediated through the activation of the A2AR/cAMP/PKA pathway, which involves its interactions with both TRAX and A2AR.

Background Fat grafting has been extensively used in plastic surgery practice, yet unstable retention in the recipient site remains a significant clinical challenge. The limited tolerance of injected adipose tissue to ischemia has prompted strategies aiming at timely enhancing the vascularity of the grafted fat. Various modified fat graft preparations have been used, and the mechanically processed tissue stromal vascular fraction (tSVF) derived from fat tissue has garnered considerable interest for enhancing rate of fat graft retention. Further enhancement of the graft retention and quality through supplements to tSVF is worthy of investigation. Methods The arteriovenous (AV) shunt in rats has been used to evaluate tSVF in vivo. We employed this animal model to investigate the regenerative potential of glycolipid Globo H Ceramide (GHCer) added to tSVF isolated from male Lewis rats. Sixty-two rats divided into four groups were studied. Study parameters included gene expression of vascular endothelial growth factor A (VEGFA) and fatty acid binding protein 4 (FABP4), percentages of the CD45 − CD31 + endothelial cell, fat tissue retention and fibrotic changes. In vitro studies on adipose-derived mesenchymal stromal cells (AD-MSCs) included angiogenesis by tube formation assay and adipogenesis. Results The addition of GHCer resulted in superior retention of the tSVF grafts at one-, two-, and eight-week post-grafting ( p  < 0.05). Elevated expression VEGFA was observed from one week ( p  < 0.05), followed by FABP4 at two weeks post-grafting in the tSVF + GHCer grafts ( p  < 0.01). After eight weeks, the numbers of CD45 − CD31 + endothelial cells and adipocytes were significantly increased in the tSVF + GHCer grafts ( p  < 0.01), while collagen deposition was reduced ( p  < 0.05). Given that GHCer potentially exerted its effects on tSVF through AD-MSCs within, we performed in vitro studies and demonstrated that GHCer promoted AD-MSC differentiation into neovessels ( p  < 0.05) and adipocytes ( p  < 0.001). Conclusions Supplementing GHCer to tSVF effectively reduced fat reabsorption and fibrotic changes of the grafts, while enhancing angiogenesis and adipogenesis, potentially through facilitating AD-MSC differentiation within tSVF. These findings support the potential clinical application of GHCer to enhance the stability and long-term outcomes of fat grafting procedures. Trial registration Not applicable. Clinical trial number Not applicable.

A sugar-lipid molecule called OAcGD2 is a novel marker for breast cancer stem cells. Treatment with anti-OAcGD2 mAb8B6 may have superior anticancer efficacy by targeting cancer stem cells, thereby reducing metastasis and recurrence of cancer. Cancer stem cells (CSCs) that drive tumor progression and disease recurrence are rare subsets of tumor cells. CSCs are relatively resistant to conventional chemotherapy and radiotherapy. Eradication of CSCs is thus essential to achieve durable responses. GD2 was reported to be a CSC marker in human triple-negative breast cancer, and anti-GD2 immunotherapy showed reduced tumor growth in cell lines. Using a specific anti-OAcGD2 antibody, mAb8D6, we set out to determine whether OAcGD2 cells exhibit stem cell properties and mAb8D6 can inhibit tumor growth by targeting OAcGD2 CSCs. OAcGD2 expression in patient-derived xenografts (PDXs) of breast cancer was determined by flow cytometric analyses using mAb8D6. The stemness of OAcGD2 cells isolated by sorting and the effects of mAb8B6 were assessed by CSC growth and mammosphere formation and tumor growth using PDX models. We found that the OAcGD2 expression levels in six PDXs of various molecular subtypes of breast cancer highly correlated with their previously defined CSC markers in these PDXs. The sorted OAcGD2 cells displayed a greater capacity for mammosphere formation and tumor initiation than OAcGD2 cells. In addition, the majority of OAcGD2 cells were aldehyde dehydrogenase (ALDH ) or CD44 CD24 , the known CSC markers in breast cancer. Treatment of PDXs-bearing mice with mAb8B6, but not doxorubicin, suppressed the tumor growth, along with reduced CSCs as assessed by CSC markers and tumorigenicity. , mAb8B6 suppressed proliferation and mammosphere formation and induced apoptosis of OAcGD2 breast cancer cells harvested from PDXs, in a dose-dependent manner. Finally, administration of mAb8B6 dramatically suppressed tumor growth of OAcGD2 breast CSCs (BCSCs) with complete tumor abrogation in 3/6 mice. OAcGD2 is a novel marker for CSC in various subtypes of breast cancer. Anti-OAcGD2 mAb8B6 directly eradicated OAcGD2 cells and reduced tumor growth in PDX model. Our data demonstrate the potential of mAb8B6 as a promising immunotherapeutic agent to target BCSCs.

Malignant cells may arise from dedifferentiation of mature cells and acquire features of the progenitor cells. Definitive endoderm from which liver is derived, expresses glycosphingolipids (GSLs) such as stage-specific embryonic antigen 3 (SSEA3), Globo H, and stage-specific embryonic antigen 4 (SSEA4). Herein, we evaluated the potential prognosis value of the three GSLs and biological functions of SSEA3 in hepatocellular carcinoma (HCC). The expression of SSEA3, Globo H, and SSEA4 in tumor tissues obtained from 328 patients with resectable HCC was examined by immunohistochemistry staining. Epithelial mesenchymal transition (EMT) and their related genes were analyzed by transwell assay and qRT-PCR, respectively. Kaplan Meier survival analysis showed significantly shorter relapse-free survival (RFS) for those with higher expression of SSEA3 (p < 0.001), Globo H (p < 0.001), and SSEA4 (p = 0.005) and worse overall survival (OS) for those with high expression of either SSEA3 (p < 0.001) or SSEA4 (p = 0.01). Furthermore, multivariable Cox regression analysis identified the SSEA3 as an independent predictor for RFS (HR: 2.68, 95% CI: 1.93-3.72, p < 0.001) and OS (HR: 2.99, 95% CI: 1.81-4.96, p < 0.001) in HCC. Additionally, SSEA3-ceramide enhanced the EMT of HCC cells, as reflected by its ability to increase migration, invasion and upregulate the expression of CDH2, vimentin, fibronectin, and MMP2, along with ZEB1. Moreover, ZEB1 silencing abrogated the EMT-enhancing effects of SSEA3-ceramide. Higher expression of SSEA3 was an independent predictor for RFS and OS in HCC and promoted EMT of HCC via upregulation of ZEB1.

The glycolipid α-galactosylceramide (α-GalCer), when presented by CD1d, can modulate the immune system through the activation of natural killer T (NKT) cells. Previously, we synthesized over 30 analogs of α-GalCer and identified a compound, C34, which features two phenyl rings on the acyl chain. C34 exhibited the most potent NKT-stimulating activities, characterized by strong Th1-biased cytokines and potent anti-tumor effects in several murine tumor models. Importantly, unlike α-GalCer, C34 did not induce NKT cell anergy. Despite these promising results, the clinical application of C34 is limited by its poor aqueous solubility. PEGylation enhances the solubility of hydrophobic drugs, and numerous PEGylated drugs have received clinical approval. Consequently, we assessed the biological activity of PEGylated C34 in this study. Murine NK1.2 cells were cultured with A20-CD1d cells in the presence of either PEGylated lipid nanocarriers encapsulating C34 (PLN-C34) or C34 dissolved in DMSO to determine IL-2 production via ELISA. C57BL/6 mice were i.v. injected with C34 or PLN-C34 to examine cytokine profiles and immune cell populations using luminex and flow cytometry, respectively. The anticancer effects of C34 and PLN-C34 were evaluated in mice bearing TC-1 lung cancer and B16 melanoma tumors. Additionally, human PBMCs were cultured with C34 or PLN-C34 to measure cytokine production through luminex. PLN-C34 demonstrated a comparable capacity to C34 in activating the NKT cell line in vitro and inducing various cytokines in vivo. Furthermore, treatment with either PLN-C34 or C34 significantly prolonged the survival of TC-1- and B16F10-bearing mice to a similar extent. Additionally, PLN-C34 effectively stimulated cytokine responses in human NKT cells, comparable to those induced by C34. These findings demonstrate that the newly formulated PLN-C34 retains NKT-stimulatory activity and anti-cancer efficacy of C34, supporting the potential of PLN as a solvent for C34 for further development in cancer therapy.

Multi-modal large language models (MLLMs) have transformed the landscape of modern healthcare, with automated radiology report generation (RRG) emerging as a cutting-edge application. While 2D MLLM-based RRG has been well established, its utility for 3D medical images remains largely unexplored. In this regard, we curate the 3D-BrainCT dataset (18,885 text-scan pairs) and develop BrainGPT, a clinically visual instruction-tuned (CVIT) model designed for 3D CT RRG. While we notice that the traditional LLM metrics failed to gauge the diagnostic quality of the RRG, we propose feature-oriented radiology task evaluation (FORTE), an evaluation scheme that captures the clinical essence of the generated reports. Here we show that BrainGPT achieves an average FORTE F1-score of 0.71 (degree = 0.661; landmark = 0.706; feature = 0.693, and impression = 0.779) and 74% of BrainGPT-generated reports were indistinguishable from human-written ground truth in a Turing -like test. Together, our work establishes a comprehensive framework encompassing dataset curation, anatomy-aware model fine-tuning, and the development of robust evaluation metrics for the RRG. By sharing our experience in 3D MLLM-based RRG, we aim to accelerate the expedition in human-machine collaboration for next-generation healthcare. Multimodal large language models (MLLMs) hold promise for a range of medical applications. Here, the authors use MLLMs for 3D brain CT radiology report generation, demonstrating that combining anatomy-aware model fine-tuning with robust evaluation metrics establishes a comprehensive and effective framework.

Background: Mesenchymal stem cells (MSCs) hold promise for cell-based therapy, yet the sourcing, quality, and invasive methods of MSCs impede their mass production and quality control. Induced pluripotent stem cell (iPSC)-derived MSCs (iMSCs) can be infinitely expanded, providing advantages over conventional MSCs in terms of meeting unmet clinical demands. Methods: The potential of MSC therapy for Leber’s hereditary optic neuropathy (LHON) remains uncertain. In this study, we used HLA-homozygous induced pluripotent stem cells to generate iMSCs using a defined protocol, and we examined their therapeutic potential in rotenone-induced LHON-like models in vitro and in vivo. Results: The iMSCs did not cause any tumorigenic incidence or inflammation-related lesions after intravitreal transplantation, and they remained viable for at least nine days in the mouse recipient’s eyes. In addition, iMSCs exhibited significant efficacy in safeguarding retinal ganglion cells (RGCs) from rotenone-induced cytotoxicity in vitro, and they ameliorated CGL+IPL layer thinning and RGC loss in vivo. Optical coherence tomography (OCT) and an electroretinogram demonstrated that iMSCs not only prevented RGC loss and impairments to the retinal architecture, but they also improved retinal electrophysiology performance. Conclusion: The generation of iMSCs via the HLA homozygosity of iPSCs offers a compelling avenue for overcoming the current limitations of MSC-based therapies. The results underscore the potential of iMSCs when addressing retinal disorders, and they highlight their clinical significance, offering renewed hope for individuals affected by LHON and other inherited retinal conditions.

Multi-modal large language models (MLLMs) have been given free rein to explore exciting medical applications with a primary focus on radiology report generation. Nevertheless, the preliminary success in 2D radiology captioning is incompetent to reflect the real-world diagnostic challenge in the volumetric 3D anatomy. To mitigate three crucial limitation aspects in the existing literature, including (1) data complexity, (2) model capacity, and (3) evaluation metric fidelity, we collected an 18,885 text-scan pairs 3D-BrainCT dataset and applied clinical visual instruction tuning (CVIT) to train BrainGPT models to generate radiology-adherent 3D brain CT reports. Statistically, our BrainGPT scored BLEU-1 = 44.35, BLEU-4 = 20.38, METEOR = 30.13, ROUGE-L = 47.6, and CIDEr-R = 211.77 during internal testing and demonstrated an accuracy of 0.91 in captioning midline shifts on the external validation CQ500 dataset. By further inspecting the captioned report, we reported that the traditional metrics appeared to measure only the surface text similarity and failed to gauge the information density of the diagnostic purpose. To close this gap, we proposed a novel Feature-Oriented Radiology Task Evaluation (FORTE) to estimate the report's clinical relevance (lesion feature and landmarks). Notably, the BrainGPT model scored an average FORTE F1-score of 0.71 (degree=0.661; landmark=0.706; feature=0.693; impression=0.779). To demonstrate that BrainGPT models possess objective readiness to generate human-like radiology reports, we conducted a Turing test that enrolled 11 physician evaluators, and around 74% of the BrainGPT-generated captions were indistinguishable from those written by humans. Our work embodies a holistic framework that showcased the first-hand experience of curating a 3D brain CT dataset, fine-tuning anatomy-sensible language models, and proposing robust radiology evaluation metrics.

Asians have higher frequency of intracranial arterial stenosis. The present study aimed to compare the clinical features and outcomes of ischemic stroke patients with and without middle cerebral artery (MCA) stenosis, assessed by transcranial sonography (TCS), based on the Taiwan Stroke Registry (TSR). Patients with acute ischemic stroke or transient ischemic attack registered in the TSR, and received both carotid duplex and TCS assessment were categorized into those with stenosis (≥50%) and without (<50%) in the extracranial internal carotid artery (ICA) and MCA, respectively. Logistic regression analysis, Kaplan-Meier method and Cox proportional hazard model were applied to assess relevant variables between groups. Of 6003 patients, 23.3% had MCA stenosis, 10.1% ICA stenosis, and 3.9% both MCA and ICA stenosis. Patients with MCA stenosis had greater initial NIHSS, higher likelihood of stroke-in-evolution, and more severe disability than those without (all p<0.001). Patients with MCA stenosis had higher prevalence of hypertension, diabetes and hypercholesterolemia. Patients with combined MCA and extracranial ICA stenosis had even higher NIHSS, worse functional outcome, higher risk of stroke recurrence or death (hazard ratio, 2.204; 95% confidence intervals, 1.440-3.374; p<0.001) at 3 months after stroke than those without MCA stenosis. In conclusion, MCA stenosis was more prevalent than extracranial ICA stenosis in ischemic stroke patients in Taiwan. Patients with MCA stenosis, especially combined extracranial ICA stenosis, had more severe neurological deficit and worse outcome.

Multi-modal large language models (MLLMs) have been given free rein to explore exciting medical applications with a primary focus on radiology report generation. Nevertheless, the preliminary success in 2D radiology captioning is incompetent to reflect the real-world diagnostic challenge in the volumetric 3D anatomy. To mitigate three crucial limitation aspects in the existing literature, including (1) data complexity, (2) model capacity, and (3) evaluation metric fidelity, we collected an 18,885 text-scan pairs 3D-BrainCT dataset and applied clinical visual instruction tuning (CVIT) to train BrainGPT models to generate radiology-adherent 3D brain CT reports. Statistically, our BrainGPT scored BLEU-1 = 44.35, BLEU-4 = 20.38, METEOR = 30.13, ROUGE-L = 47.6, and CIDEr-R = 211.77 during internal testing and demonstrated an accuracy of 0.91 in captioning midline shifts on the external validation CQ500 dataset. By further inspecting the captioned report, we reported that the traditional metrics appeared to measure only the surface text similarity and failed to gauge the information density of the diagnostic purpose. To close this gap, we proposed a novel Feature-Oriented Radiology Task Evaluation (FORTE) to estimate the report's clinical relevance (lesion feature and landmarks). Notably, the BrainGPT model scored an average FORTE F1-score of 0.71 (degree=0.661; landmark=0.706; feature=0.693; impression=0.779). To demonstrate that BrainGPT models possess objective readiness to generate human-like radiology reports, we conducted a Turing test that enrolled 11 physician evaluators, and around 74% of the BrainGPT-generated captions were indistinguishable from those written by humans. Our work embodies a holistic framework that showcased the first-hand experience of curating a 3D brain CT dataset, fine-tuning anatomy-sensible language models, and proposing robust radiology evaluation metrics.