A fine-grained recognition technique for identifying Chinese food images.

As a crucial area of research in the field of computer vision, food recognition technology has become a core technology in many food-related fields, such as unmanned restaurants and food nutrition analysis, which are closely related to our healthy lives. Obtaining accurate classification results is the most important task in food recognition. Food classification is a fine-grained recognition process, which involves extracting features from a group of objects with similar appearances and accurately classifying them into different categories. In a such usage environment, the network is required to not only overview the overall image, but also capture the subtle details within it. In addition, since Chinese food images have unique texture features, the model needs to extract texture information from the image. However, existing CNN methods have not focused on and processed this information. To classify food as accurately as possible, this paper introduces the Laplace pyramid into the convolution layer and proposes a bilinear network that can perceive image texture features and multi-scale features (LMB-Net). The proposed model was evaluated on a public dataset, and the results demonstrate that LMB-Net achieves state-of-the-art classification performance.

Cullin-associated and neddylation-dissociated protein 1 (CAND1) alleviates NAFLD by reducing ubiquitinated degradation of ACAA2.

Nonalcoholic fatty liver disease (NAFLD) is the most common liver disorder with high morbidity and mortality. The current study aims to explore the role of Cullin-associated and neddylation-dissociated protein 1 (CAND1) in the development of NAFLD and the underlying mechanisms. CAND1 is reduced in the liver of NAFLD male patients and high fat diet (HFD)-fed male mice. CAND1 alleviates palmitate (PA) induced lipid accumulation in vitro. Hepatocyte-specific knockout of CAND1 exacerbates HFD-induced liver injury in HFD-fed male mice, while hepatocyte-specific knockin of CAND1 ameliorates these pathological changes. Mechanistically, deficiency of CAND1 enhances the assembly of Cullin1, F-box only protein 42 (FBXO42) and acetyl-CoA acyltransferase 2 (ACAA2) complexes, and thus promotes the ubiquitinated degradation of ACAA2. ACAA2 overexpression abolishes the exacerbated effects of CAND1 deficiency on NAFLD. Additionally, androgen receptor binds to the -187 to -2000 promoter region of CAND1. Collectively, CAND1 mitigates NAFLD by inhibiting Cullin1/FBXO42 mediated ACAA2 degradation.

Interdependent Nuclear Co-Trafficking of ASPP1 and p53 Aggravates Cardiac Ischemia/Reperfusion Injury.

OBJECTIVE: ASPP1 (apoptosis stimulating of p53 protein 1) is critical in regulating cell apoptosis as a cofactor of p53 to promote its transcriptional activity in the nucleus. However, whether cytoplasmic ASPP1 affects p53 nuclear trafficking and its role in cardiac diseases remains unknown. This study aims to explore the mechanism by which ASPP1 modulates p53 nuclear trafficking and the subsequent contribution to cardiac ischemia/reperfusion (I/R) injury. METHODS AND RESULTS: The immunofluorescent staining showed that under normal condition ASPP1 and p53 colocalized in the cytoplasm of neonatal mouse ventricular cardiomyocytes, while they were both upregulated and translocated to the nuclei upon hypoxia/reoxygenation treatment. The nuclear translocation of ASPP1 and p53 was interdependent, as knockdown of either ASPP1 or p53 attenuated nuclear translocation of the other one. Inhibition of importin-ß1 resulted in the cytoplasmic sequestration of both p53 and ASPP1 in neonatal mouse ventricular cardiomyocytes with hypoxia/reoxygenation stimulation. Overexpression of ASPP1 potentiated, whereas knockdown of ASPP1 inhibited the expression of Bax (Bcl2-associated X), PUMA (p53 upregulated modulator of apoptosis), and Noxa, direct apoptosis-associated targets of p53. ASPP1 was also increased in the I/R myocardium. Cardiomyocyte-specific transgenic overexpression of ASPP1 aggravated I/R injury as indicated by increased infarct size and impaired cardiac function. Conversely, knockout of ASPP1 mitigated cardiac I/R injury. The same qualitative data were observed in neonatal mouse ventricular cardiomyocytes exposed to hypoxia/reoxygenation injury. Furthermore, inhibition of p53 significantly blunted the proapoptotic activity and detrimental effects of ASPP1 both in vitro and in vivo. CONCLUSIONS: Binding of ASPP1 to p53 triggers their nuclear cotranslocation via importin-ß1 that eventually exacerbates cardiac I/R injury. The findings imply that interfering the expression of ASPP1 or the interaction between ASPP1 and p53 to block their nuclear trafficking represents an important therapeutic strategy for cardiac I/R injury.

Automatic tongue image quality assessment using a multi-task deep learning model.

The quality of tongue images has a significant influence on the performance of tongue diagnosis in Chinese medicine. During the acquisition process, the quality of the tongue image is easily affected by factors such as the illumination, camera parameters, and tongue extension of the subject. To ensure that the quality of the collected images meet the diagnostic criteria of traditional Chinese Medicine practitioners, we propose a deep learning model to evaluate the quality of tongue images. First, we acquired the tongue images of the patients under different lighting conditions, exposures, and tongue extension conditions using the inspection instrument, and experienced Chinese physicians manually screened them into high-quality and unqualified tongue datasets. We then designed a multi-task deep learning network to classify and evaluate the quality of tongue images by adding tongue segmentation as an auxiliary task, as the two tasks are related and can promote each other. Finally, we adaptively designed different task weight coefficients of a multi-task network to obtain better tongue image quality assessment (IQA) performance, as the two tasks have relatively different contributions in the loss weighting scheme. Experimental results show that the proposed method is superior to the traditional deep learning tongue IQA method, and as an additional task of the network, it can output the tongue segmentation area, which provides convenience for follow-up clinical tongue diagnosis. In addition, we used network visualization to verify the effectiveness of the proposed method qualitatively.

Cytoplasmic sequestration of p53 by lncRNA-CIRPILalleviates myocardial ischemia/reperfusion injury.

Myocardial ischemia/reperfusion (MI/R) injury is a pathological process that seriously affects the health of patients with coronary artery disease. Long non-coding RNAs (lncRNAs) represents a new class of regulators of diverse biological processes and disease conditions, the study aims to discover the pivotal lncRNA in MI/R injury. The microarray screening identifies a down-regulated heart-enriched lncRNA-CIRPIL (Cardiac ischemia reperfusion associated p53 interacting lncRNA, lncCIRPIL) from the hearts of I/R mice. LncCIRPIL inhibits apoptosis of cultured cardiomyocytes exposed to anoxia/reoxygenation (A/R). Cardiac-specific transgenic overexpression of lncCIRPIL alleviates I/R injury in mice, while knockout of lncCIRPIL exacerbates cardiac I/R injury. LncCIRPIL locates in the cytoplasm and physically interacts with p53, which leads to the cytoplasmic sequestration and the acceleration of ubiquitin-mediated degradation of p53 triggered by E3 ligases CHIP, COP1 and MDM2. p53 overexpression abrogates the protective effects of lncCIRPIL. Notably, the human fragment of conserved lncCIRPIL mimics the protective effects of the full-length lncCIRPIL on cultured human AC16 cells. Collectively, lncCIRPIL exerts its cardioprotective action via sequestering p53 in the cytoplasm and facilitating its ubiquitin-mediated degradation. The study highlights a unique mechanism in p53 signal pathway and broadens our understanding of the molecular mechanisms of MI/R injury.

Weakly Supervised Deep Learning for Tooth-Marked Tongue Recognition.

The recognition of tooth-marked tongues has important value for clinical diagnosis of traditional Chinese medicine. Tooth-marked tongue is often related to spleen deficiency, cold dampness, sputum, effusion, and blood stasis. The clinical manifestations of patients with tooth-marked tongue include loss of appetite, borborygmus, gastric distention, and loose stool. Traditional clinical tooth-marked tongue recognition is conducted subjectively based on the doctor's visual observation, and its performance is affected by the doctor's subjectivity, experience, and environmental lighting changes. In addition, the tooth marks typically have various shapes and colors on the tongue, which make it very challenging for doctors to identify tooth marks. The existing methods based on deep learning have made great progress for tooth-marked tongue recognition, but there are still shortcomings such as requiring a large amount of manual labeling of tooth marks, inability to detect and locate the tooth marks, and not conducive to clinical diagnosis and interpretation. In this study, we propose an end-to-end deep neural network for tooth-marked tongue recognition based on weakly supervised learning. Note that the deep neural network only requires image-level annotations of tooth-marked or non-tooth marked tongues. In this method, a deep neural network is trained to classify tooth-marked tongues with the image-level annotations. Then, a weakly supervised tooth-mark detection network (WSTDN) as an architecture variant of the pre-trained deep neural network is proposed for the tooth-marked region detection. Finally, the WSTDN is re-trained and fine-tuned using only the image-level annotations to simultaneously realize the classification of the tooth-marked tongue and the positioning of the tooth-marked region. Experimental results of clinical tongue images demonstrate the superiority of the proposed method compared with previously reported deep learning methods for tooth-marked tongue recognition. The proposed tooth-marked tongue recognition model may provide important syndrome diagnosis and efficacy evaluation methods, and contribute to the understanding of ethnopharmacological mechanisms.

Adaptive Multimodal Fusion With Attention Guided Deep Supervision Net for Grading Hepatocellular Carcinoma.

Multimodal medical imaging plays a crucial role in the diagnosis and characterization of lesions. However, challenges remain in lesion characterization based on multimodal feature fusion. First, current fusion methods have not thoroughly studied the relative importance of characterization modals. In addition, multimodal feature fusion cannot provide the contribution of different modal information to inform critical decision-making. In this study, we propose an adaptive multimodal fusion method with an attention-guided deep supervision net for grading hepatocellular carcinoma (HCC). Specifically, our proposed framework comprises two modules: attention-based adaptive feature fusion and attention-guided deep supervision net. The former uses the attention mechanism at the feature fusion level to generate weights for adaptive feature concatenation and balances the importance of features among various modals. The latter uses the weight generated by the attention mechanism as the weight coefficient of each loss to balance the contribution of the corresponding modal to the total loss function. The experimental results of grading clinical HCC with contrast-enhanced MR demonstrated the effectiveness of the proposed method. A significant performance improvement was achieved compared with existing fusion methods. In addition, the weight coefficient of attention in multimodal fusion has demonstrated great significance in clinical interpretation.

Attention guided discriminative feature learning and adaptive fusion for grading hepatocellular carcinoma with Contrast-enhanced MR.

Multimodality medical imaging has played a significant role in lesion diagnosis and characterization. However, there are remaining challenges in the procedure of multimodality feature fusion based lesion characterization. First, large inter-modality variations make it difficult to harness the complementary information between modalities for better characterization. Subsequently, large intra-class and small inter-class variations due to the heterogeneity of neoplasm makes the classification more challenging. Finally, the relative importance of modalities for the characterization has not been thoroughly investigated, easily resulting in non-optimal fusion performance. In this study, we propose an attention guided discriminative and adaptive fusion (AGDAF) method based on deep learning architecture to address above three problems. Specifically, we first design a novel cross-modal intra- and inter-attention module to focus on learning both the intra-modality relations and inter-modality relations. Then, we introduce a discriminative feature learning loss to reduce the distance of features in the same class and increase the distance of features in different classes of neoplasm in single modalities. Finally, we design an adaptive weighting strategy to increase the contribution of modalities with relatively lower loss values and reduce the impact of modalities with large loss values for the final loss function. Experimental results of grading clinical hepatocellular carcinoma demonstrate that the proposed method significantly outperforms the previously reported multimodality feature fusion methods. In addition, ablation study also demonstrates the effectiveness of the proposed cross-modal intra- and inter-attention module, discriminative module, and adaptive weight adjustment module for multimodality feature fusion in lesion characterization.

Apolipoprotein A-I mimetic peptide inhibits atherosclerosis by increasing tetrahydrobiopterin via regulation of GTP-cyclohydrolase 1 and reducing uncoupled endothelial nitric oxide synthase activity.

BACKGROUND AND AIMS: The apolipoprotein A-I mimetic peptide D-4F, among its anti-atherosclerotic effects, improves vasodilation through mechanisms not fully elucidated yet. METHODS: Low-density lipoprotein (LDL) receptor null (LDLr-/-) mice were fed Western diet with or without D-4F. We then measured atherosclerotic lesion formation, endothelial nitric oxide synthase (eNOS) phosphorylation and its association with heat shock protein 90 (HSP90), nitric oxide (NO) and superoxide anion (O2•-) production, and tetrahydrobiopterin (BH4) and GTP-cyclohydrolase 1 (GCH-1) concentration in the aorta. Human umbilical vein endothelial cells (HUVECs) and aortas were treated with oxidized LDL (oxLDL) with or without D-4F; subsequently, BH4 and GCH-1 concentration, NO and O2•- production, eNOS association with HSP90, and endothelium-dependent vasodilation were measured. RESULTS: Unexpectedly, eNOS phosphorylation, eNOS-HSP90 association, and O2•- production were increased, whereas BH4 and GCH-1 concentration and NO production were reduced in atherosclerosis. D-4F significantly inhibited atherosclerosis, eNOS phosphorylation, eNOS-HSP90 association, and O2•- generation but increased NO production and BH4 and GCH-1 concentration. OxLDL reduced NO production and BH4 and GCH-1 concentration but enhanced O2•- generation and eNOS association with HSP90, and impaired endothelium-dependent vasodilation. D-4F inhibited the overall effects of oxLDL. CONCLUSIONS: Hypercholesterolemia enhanced uncoupled eNOS activity by decreasing GCH-1 concentration, thereby reducing BH4 levels. D-4F reduced uncoupled eNOS activity by increasing BH4 levels through GCH-1 expression and decreasing eNOS phosphorylation and eNOS-HSP90 association. Our findings elucidate a novel mechanism by which hypercholesterolemia induces atherosclerosis and D-4F inhibits it, providing a potential therapeutic approach.

Simvastatin inhibits POVPC-mediated induction of endothelial-to-mesenchymal cell transition.

Endothelial-to-mesenchymal transition (EndMT), the process by which an endothelial cell (EC) undergoes a series of molecular events that result in a mesenchymal cell phenotype, plays an important role in atherosclerosis. 1-Palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphocholine (POVPC), derived from the oxidation of 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphatidylcholine, is a proinflammatory lipid found in atherosclerotic lesions. Whether POVPC promotes EndMT and how simvastatin influences POVPC-mediated EndMT remains unclear. Here, we treated human umbilical vein ECs with POVPC, simvastatin, or both, and determined their effect on EC viability, morphology, tube formation, proliferation, and generation of NO and superoxide anion (O2•-). Expression of specific endothelial and mesenchymal markers was detected by immunofluorescence and immunoblotting. POVPC did not affect EC viability but altered cellular morphology from cobblestone-like ECs to a spindle-like mesenchymal cell morphology. POVPC increased O2- generation and expression of alpha-smooth muscle actin, vimentin, Snail-1, Twist-1, transforming growth factor-beta (TGF-ß), TGF-ß receptor II, p-Smad2/3, and Smad2/3. POVPC also decreased NO production and expression of CD31 and endothelial NO synthase. Simvastatin inhibited POVPC-mediated effects on cellular morphology, production of O2•- and NO, and expression of specific endothelial and mesenchymal markers. These data demonstrate that POVPC induces EndMT by increasing oxidative stress, which stimulates TGF-ß/Smad signaling, leading to Snail-1 and Twist-1 activation. Simvastatin inhibited POVPC-induced EndMT by decreasing oxidative stress, suppressing TGF-ß/Smad signaling, and inactivating Snail-1 and Twist-1. Our findings reveal a novel mechanism of atherosclerosis that can be inhibited by simvastatin.

The long noncoding RNA lncCIRBIL disrupts the nuclear translocation of Bclaf1 alleviating cardiac ischemia-reperfusion injury.

Cardiac ischemia-reperfusion (I/R) injury is a pathological process resulting in cardiomyocyte death. The present study aims to evaluate the role of the long noncoding RNA Cardiac Injury-Related Bclaf1-Inhibiting LncRNA (lncCIRBIL) on cardiac I/R injury and delineate its mechanism of action. The level of lncCIRBIL is reduced in I/R hearts. Cardiomyocyte-specific transgenic overexpression of lncCIRBIL reduces infarct area following I/R injury. Knockout of lncCIRBIL in mice exacerbates cardiac I/R injury. Qualitatively, the same results are observed in vitro. LncCIRBIL directly binds to BCL2-associated transcription factor 1 (Bclaf1), to inhibit its nuclear translocation. Cardiomyocyte-specific transgenic overexpression of Bclaf1 worsens, while partial knockout of Bclaf1 mitigates cardiac I/R injury. Meanwhile, partial knockout of Bclaf1 abrogates the detrimental effects of lncCIRBIL knockout on cardiac I/R injury. Collectively, the protective effect of lncCIRBIL on I/R injury is accomplished by inhibiting the nuclear translocation of Bclaf1. LncCIRBIL and Bclaf1 are potential therapeutic targets for ischemic cardiac disease.

Circulating extracellular vesicles from patients with valvular heart disease induce neutrophil chemotaxis via FOXO3a and the inhibiting role of dexmedetomidine.

We previously demonstrated that circulating extracellular vesicles (EVs) from patients with valvular heart disease (VHD; vEVs) contain inflammatory components and inhibit endothelium-dependent vasodilation. Neutrophil chemotaxis plays a key role in renal dysfunction, and dexmedetomidine (DEX) can reduce renal dysfunction in cardiac surgery. However, the roles of vEVs in neutrophil chemotaxis and effects of DEX on vEVs are unknown. Here, we investigated the impact of vEVs on neutrophil chemotaxis in kidneys and the influence of DEX on vEVs. Circulating EVs were isolated from healthy subjects and patients with VHD. The effects of EVs on chemokine generation, forkhead box protein O3a (FOXO3a) pathway activation and neutrophil chemotaxis on cultured human umbilical vein endothelial cells (HUVECs) and kidneys in mice and the influence of DEX on EVs were detected. vEVs increased FOXO3a expression, decreased phosphorylation of Akt and FOXO3a, promoted FOXO3a nuclear translocation, and activated the FOXO3a signaling pathway in vitro. DEX pretreatment reduced vEV-induced CXCL4 and CCL5 expression and neutrophil chemotaxis in cultured HUVECs via the FOXO3a signaling pathway. vEVs were also found to suppress Akt phosphorylation and activate FOXO3a signaling to increase plasma levels of CXCL4 and CCL5 and neutrophil accumulation in kidney. The overall mechanism was inhibited in vivo with DEX pretreatment. Our data demonstrated that vEVs induced CXCL4-CCL5 to stimulate neutrophil infiltration in kidney, which can be inhibited by DEX via the FOXO3a signaling. Our findings reveal a unique mechanism involving vEVs in inducing neutrophils chemotaxis and may provide a novel basis for using DEX in reducing renal dysfunction in valvular heart surgery.

Valsartan reduced the vulnerability to atrial fibrillation by preventing action potential prolongation and conduction slowing in castrated male mice.

INTRODUCTION: Deficiency of testosterone was associated with the susceptibility of atrial fibrillation (AF). Angiotensin-II (AngII) receptor antagonists were shown to reduce AF by improving atrial electrical remodeling. This study investigated the effects and mechanism of valsartan, an AngII receptor antagonist, on the susceptibility to AF with testosterone deficiency. METHODS AND RESULTS: Five-week-old male ICR mice were castrated and valsartan was administered orally (50 mg/kg/d). High-frequency electrical stimulation method was used to induce atrial arrhythmia. Patch-clamp technique was used for recording action potential duration (APD), transient outward potassium current ( I to ), sustained outward potassium current ( I ksus ), and late sodium current ( I Na-L ). Optical mapping technique was used to examine atrial conduction velocity (CV). The expression of connexin40 (Cx40) and Cx43 were detected by Western blot analysis. The occurrence rate of AF was significantly increased in castrated mice and APDs measured at 50% and 90% repolarization were markedly prolonged in castrated mice than controls, which were alleviated by the administration of valsartan. Valsartan suppressed the increase of INa-L and rescued the reduction of Ito and Iksus in castrated mice. The left atrial CV in castrated mice was decreased and the expression of Cx43 reduced than controls, which were restored after valsartan treatment. CONCLUSIONS: Valsartan reduced the susceptibility of AF in castrated mice, which may be related to the inhibition of action potential prolongation and improvement of atrial conduction impairment. This study indicates that valsartan may represent a useful agent for the prevention of AF pathogenesis in elderly male patients.

The oxidized phospholipid POVPC impairs endothelial function and vasodilation via uncoupling endothelial nitric oxide synthase.

Endothelial dysfunction is an early stage of atherosclerosis. We recently have shown that 25-hydroxycholesterol found in atherosclerotic lesions could impair endothelial function and vasodilation by uncoupling and inhibiting endothelial nitric oxide synthase (eNOS). 1-Palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphocholine (POVPC), the oxidation product of oxidized low-density lipoprotein, is another proinflammatory lipid and has also been found in atherosclerotic lesions. However, whether POVPC promotes atherosclerosis like 25-hydroxycholesterol remains unclear. The purpose of this study was to explore the effects of POVPC on endothelial function and vasodilation. Human umbilical vein endothelial cells (HUVECs) were incubated with POVPC. Endothelial cell proliferation, migration and tube formation were measured. Nitric oxide (NO) production and superoxide anion generation (O2-) were determined. The expression and phosphorylation of endothelial nitric oxide synthase (eNOS), AKT, PKC-ßII and P70S6K as well as the association of eNOS and heat shock protein 90 (HSP90) were detected by immunoblotting and immunoprecipitation. Endothelial cell apoptosis was monitored by TUNEL staining. The expression of Bcl-2, Bax, and Cleaved Caspase 3 were detected by immunoblotting. Finally, aortic ring from C57BL6 mice were isolated and treated with POVPC and the endothelium-dependent vasodilation was evaluated. POVPC significantly inhibited HUVECs proliferation, migration, tube formation, decreased NO production but increased O2- generation. POVPC inhibited the phosphorylation of Akt and eNOS at Ser1177, increased activation of PKC-ßII, P70S6K and the phosphorylation of eNOS at Thr495, reduced the association of HSP90 with eNOS. Meanwhile, POVPC induced endothelial cell apoptosis by inhibiting Bcl-2 expression, increasing Bax and cleaved caspase-3 expressions as well as caspase-3 activity and impaired endothelium-dependent vasodilation. These data demonstrated that POVPC impaired endothelial function by uncoupling and inhibiting eNOS as well as by inducing endothelial cell apoptosis. Therefore, POVPC may play an important role in the development of atherosclerosis and may be considered as a potential therapeutic target for atherosclerosis.