Rare bioparticle detection via deep metric learning

Recent deep neural networks have shown superb performance in analyzing bioimages for disease diagnosis and bioparticle classification. Conventional deep neural networks use simple classifiers such as SoftMax to obtain highly accurate results. However, they have limitations in many practical applications that require both low false alarm rate and high recovery rate, e.g., rare bioparticle detection, in which the representative image data is hard to collect, the training data is imbalanced, and the input images in inference time could be different from the training images. Deep metric learning offers a better generatability by using distance information to model the similarity of the images and learning function maps from image pixels to a latent space, playing a vital role in rare object detection. In this paper, we propose a robust model based on a deep metric neural network for rare bioparticle (Cryptosporidium or Giardia) detection in drinking water. Experimental results showed that the deep metric neural network achieved a high accuracy of 99.86% in classification, 98.89% in precision rate, 99.16% in recall rate and zero false alarm rate. The reported model empowers imaging flow cytometry with capabilities of biomedical diagnosis, environmental monitoring, and other biosensing applications.

Conventional deep neural networks use simple classifiers to obtain highly accurate results. However, they have limitations in practical applications. This study demonstrates a robust deep metric neural network model for rare bioparticle detection.     

Landscape of surfaceome and endocytome in human glioma is divergent and depends on cellular spatial organization

Therapeutic strategies directed at the tumor surfaceome (TS), including checkpoint inhibitor blocking antibodies, antibody drug conjugates (ADCs), and chimeric antigen receptor T (CAR-T) cells, provide a new armament to fight cancer. However, a remaining bottleneck is the lack of strategies to comprehensively interrogate patient tumors for potential TS targets. Here, we have developed a platform (tumor surfaceome mapping [TS-MAP]) integrated with a newly curated TS classifier (SURFME) that allows profiling of primary 3D cultures and intact patient glioma tumors with preserved tissue architecture. Moreover, TS-MAP specifically identifies proteins capable of endocytosis as tractable targets for ADCs and other modalities requiring toxic payload internalization. In high-grade gliomas that remain among the most aggressive forms of cancer, we show that cellular spatial organization (2D vs. 3D) fundamentally transforms the surfaceome and endocytome (e.g., integrins, proteoglycans, semaphorins, and cancer stem cell markers) with general implications for target screening approaches, as exemplified by an ADC targeting EGFR. The TS-MAP platform was further applied to profile the surfaceome and endocytome landscape in a cohort of freshly resected gliomas. We found a highly diverse TS repertoire between patient tumors, not directly associated with grade and histology, which highlights the need for individualized approaches. Our data provide additional layers of understanding fundamental to the future development of immunotherapy strategies, as well as procedures for proteomics-based target identification and selection. The TS-MAP platform should be widely applicable in efforts aiming at a better understanding of how to harness the TS for personalized immunotherapy.

Cell-surface proteins have a key role in drug development, and approximately two-thirds of approved human drugs target a cell-surface protein (1). Recently, tumor cell–surface proteins integrated with the plasma membrane (tumor surfaceome [TS]) have attracted considerable attention as targets for immunotherapies in cancer. Immune checkpoint-blocking antibodies (e.g., ipilimumab and nivolumab), antibody drug conjugates (ADCs, e.g., trastuzumab emtansin), radioimmunotherapy (RIT, e.g., ⁹⁰Y ibritumomab tiuxetan), and chimeric antigen receptor T (CAR-T) cells are all directed at the TS and currently revolutionize cancer treatment (2–6). With the impressive development of creative methods for antibody and T cell engineering, a remaining challenge is the lack of strategies to comprehensively map potential TS target antigens for the design of more rational, individualized treatments (7). Although advancements in DNA and RNA sequencing provide high throughput data for prediction algorithms, e.g., personalized peptide vaccine trials (8, 9), the predicted proteome derived from these platforms is not necessarily expressed and available for targeting. Moreover, proteomics-based strategies involve analysis of the bulk from disintegrated tumor tissue, resulting in loss of spatial information and limited coverage of the less abundant and hydrophobic TS proteins (10, 11). Of particular relevance, ADC, RIT, and other intracellular drug delivery strategies rely on TS targets that functionally engage in endocytic internalization (12). Clearly, despite its great targeting potential in cancer immunotherapy, the TS remains an elusive treasure for further discovery.Procedures for unbiased mapping of the TS and target identification should include specific labeling of the TS in freshly resected patient tumors with preserved tissue architecture. Enrichment of TS proteins and reduction of noise from intracellular proteins as well as abundant extracellular matrix collagens and glycoproteins would greatly improve downstream mass spectrometry analysis. Moreover, the approach should allow functional and dynamic profiling of TS internalization in an intact tissue environment. With the aim to address these challenges and to provide insight into the complexity of the TS, we have developed a versatile technology for TS mapping (TS-MAP). As proof of concept, we focused on primary brain tumors that remain among the most aggressive forms of cancer and for which attempts to conquer the most common variant, glioblastoma (GBM) (World Health Organization [WHO] grade IV) have failed so far (13). TS-MAP is compatible with spheroids from primary human stem cell–like GBM cultures, as well as mouse and patient brain tumors, and separately profiles surface resident and internalized TS proteins. Moreover, a TS classifier (SURFME) was curated for filtering and categorization of bona fide membrane proteins exposed to the extracellular space. We find significant differences in the TS between the 2D and 3D spheroid format, which underlines the importance of cellular spatial organization. In strong support of the need of individualized approaches, our findings suggest substantial intertumoral heterogeneity in the relative abundance of TS proteins in a cohort of freshly resected patient gliomas.     

The blood-brain barrier models to study apolipoprotein E genotypes in Alzheimer’s disease

Alzheimer’s disease(AD)is a neurodegenerative disease that is characterized by an age-dependent progressive decline of memory,impairment of cognitive functions and changes in personality and behavior.Despite the improvement in understanding of the mechanisms underlying the disease,AD remains an incurable complex disorder with multifaceted pathophysiology to date.Apolipoprotein E(ApoE)is the main cholesterol carrier in the brain that supports lipid transport between brain cells.The individuals carrying the APOE4 allele are known to be at increased risk of developing AD compared with those carrying the more common APOE3 allele.     

Management of apparent life-threatening events

Apparent life-threatening events in infancy (ALTE) present a common yet complex management problem for the clinician. While an ALTE generally represents a benign event, in rarer instances it may indicate a serious underlying disorder. In most circumstances patients will require admission for a short period, thus providing the opportunity to perform a systematic, thorough examination, followed by the selective use of investigations. In all situations, however, even those with an unambiguous diagnosis, follow-up must be provided to detect recurrent episodes and to monitor long-term sequelae.     

Posttraumatic Stress Disorder Is Associated With Emotional Eating

The present study investigated the relationship between posttraumatic stress disorder (PTSD) and emotional eating in a sample of medically healthy and medication‐free adults. Participants with PTSD (n = 44) and control participants free of lifetime psychiatric history (n = 49) completed a measure of emotional eating. Emotional eating is the tendency to eat or overeat in response to negative emotions. PTSD participants exhibited greater emotional eating than control participants (η² = .20) and emotional eating increased with higher PTSD symptom severity (R² = .11). Results supported the stress‐eating‐obesity model whereby emotional eating is a maladaptive response to stressors. Over time, this could lead to weight gain, particularly abdominal stores, and contribute to higher risk for comorbid medical disorders. Findings suggest the importance of future longitudinal research to understand whether emotional eating contributes to the high rates of obesity, diabetes, and heart disease in PTSD.     

A randomized phase II pharmacokinetic and pharmacodynamic study of indisulam as second-line therapy in patients with advanced non-small cell lung cancer.

PURPOSE: The primary aim of this study was to measure the objective tumor response rate following treatment with indisulam [E7070; N-(3-chloro-7-indolyl)-1,4-benzenedisulfonamide] as second-line therapy in patients with advanced non-small cell lung cancer. The secondary aims were to determine progression-free survival, to assess the safety and tolerability of indisulam, and to study its pharmacokinetic and pharmacodynamic profile. EXPERIMENTAL DESIGN: Patients were randomized to receive indisulam every 3 weeks either as a single i.v. dose of 700 mg/m(2) on day one (dx1) or 130 mg/m(2) given on days 1 to 5 inclusive as a daily infusion (dx5). All patients had previously received platinum-based chemotherapy. RESULTS: Forty-four patients were randomized. Only minor responses were seen. Myelosuppression, gastrointestinal symptoms, and lethargy were the most common toxicities and were more frequent in the dx1 arm. The pharmacokinetics of indisulam in each treatment schedule were adequately described using a previously developed population pharmacokinetic model and were mostly consistent with the results of the phase I program. Flow cytometric analysis of endobronchial and metastatic disease revealed a reduction in the fraction of cycling cells and an increase in apoptosis following indisulam compared with pretreatment levels. CONCLUSIONS: We conclude that, despite evidence of tumor-specific indisulam-induced apoptosis, neither of these treatment schedules has single-agent activity as second-line treatment of non-small cell lung cancer.