Analysis of unknown (unlabeled/mislabeled) drug products for active pharmaceutical ingredients and related substances by an international mail facility satellite laboratory equipped with rapid screening devices.

Two chemists employed a three-device rapid screening "toolkit" consisting of a handheld Raman spectrometer, transportable mass spectrometer, and portable Fourier transform infrared (FT-IR) spectrometer at an international mail facility (IMF) satellite laboratory to examine unknown (unlabeled/mislabeled) products for the presence of active pharmaceutical ingredients (APIs). Phase I of this project previously demonstrated that this toolkit was the most effective collection of instruments for identifying APIs in product types collected at IMFs during a nationwide mail blitz and Phase II of this project previously demonstrated that results generated using the toolkit during a satellite laboratory pilot program were as reliable as those generated by a full-service library when two or more of these instruments identify an API. This study (Phase III) described the results of the satellite laboratory toolkit during production mode and encompassed the period ranging from June 2021 through December 2022. During this study, a total of 858 products were examined on-site at the IMF. The satellite laboratory yielded conclusive results for 726 (84.6%) products, which were used to support regulatory action, and identified 132 (15.4%) products that required additional full-service laboratory analyses due to inconclusive results. The satellite and full-service laboratory verified/confirmed at least one API/related substance in 617 (71.9%) products. A total of 709 APIs/related substances were found in the 617 products, and 202 of these 709 compounds were unique/different. Overall, during Phases I through III of this program, 350 different substances have been identified in products collected at IMFs.

Rapid screening of 2-benzylbenzimidazole nitazene analogs in suspect counterfeit tablets using Raman, SERS, DART-TD-MS, and FT-IR.

Developing methods to rapidly screen for novel synthetic 2-benzylbenzimidazole opioids, also known as nitazenes, has become increasingly important due to their high potency. These compounds have potency comparable or exceeding that of fentanyl by up to 10 times and have been implicated in approximately 5% of all drug overdose deaths in the United States in 2021. This paper details the authenticity determination of suspect tablets and the identification of three nitazene analogs (N-pyrrolidino etonitazene, isotonitazene, and etodesnitazene) in suspect tablets seized at a mail facility using Raman and surface-enhanced Raman scattering (SERS) with handheld devices, portable Fourier transform infrared spectrometer (FT-IR), and a direct analysis in real-time ambient ionization coupled to a thermal desorption unit and a mass spectrometer (DART-TD-MS). These methods are rapid and excellent for screening opioids in suspect tablets but could not fully determine the exact structure of some of the nitazene analogs present due to spectral similarities or similar fragmentation patterns. Liquid chromatography-mass spectrometry (LC-MS) confirmed the presence of these nitazene compounds in addition to other opioids/drugs that were in trace quantities. The quantitative high-performance liquid chromatography coupled with ultraviolet (HPLC-UV) detection experiments determined that the suspect tablets contained an average of 0.817 mg of N-pyrrolidino etonitazene per tablet. The results obtained reveal that the simultaneous deployment of these complementary and orthogonal portable analytical techniques as part of a workflow allows suspect tablets to be screened and nitazene-type drugs to be identified in suspect counterfeit tablets at remote sampling sites.

Harmony: A Generic Unsupervised Approach for Disentangling Semantic Content from Parameterized Transformations.

In many real-life image analysis applications, particularly in biomedical research domains, the objects of interest undergo multiple transformations that alters their visual properties while keeping the semantic content unchanged. Disentangling images into semantic content factors and transformations can provide significant benefits into many domain-specific image analysis tasks. To this end, we propose a generic unsupervised framework, Harmony, that simultaneously and explicitly disentangles semantic content from multiple parameterized transformations. Harmony leverages a simple cross-contrastive learning framework with multiple explicitly parameterized latent representations to disentangle content from transformations. To demonstrate the efficacy of Harmony, we apply it to disentangle image semantic content from several parameterized transformations (rotation, translation, scaling, and contrast). Harmony achieves significantly improved disentanglement over the baseline models on several image datasets of diverse domains. With such disentanglement, Harmony is demonstrated to incentivize bioimage analysis research by modeling structural heterogeneity of macromolecules from cryo-ET images and learning transformation-invariant representations of protein particles from single-particle cryo-EM images. Harmony also performs very well in disentangling content from 3D transformations and can perform coarse and fast alignment of 3D cryo-ET subtomograms. Therefore, Harmony is generalizable to many other imaging domains and can potentially be extended to domains beyond imaging as well.

UNSUPERVISED DOMAIN ALIGNMENT BASED OPEN SET STRUCTURAL RECOGNITION OF MACROMOLECULES CAPTURED BY CRYO-ELECTRON TOMOGRAPHY.

Cellular cryo-Electron Tomography (cryo-ET) provides three-dimensional views of structural and spatial information of various macromolecules in cells in a near-native state. Subtomogram classification is a key step for recognizing and differentiating these macromolecular structures. In recent years, deep learning methods have been developed for high-throughput subtomogram classification tasks; however, conventional supervised deep learning methods cannot recognize macromolecular structural classes that do not exist in the training data. This imposes a major weakness since most native macromolecular structures in cells are unknown and consequently, cannot be included in the training data. Therefore, open set learning which can recognize unknown macromolecular structures is necessary for boosting the power of automatic subtomogram classification. In this paper, we propose a method called Margin-based Loss for Unsupervised Domain Alignment (MLUDA) for open set recognition problems where only a few categories of interest are shared between cross-domain data. Through extensive experiments, we demonstrate that MLUDA performs well at cross-domain open-set classification on both public datasets and medical imaging datasets. So our method is of practical importance.

End-to-end robust joint unsupervised image alignment and clustering.

Computing dense pixel-to-pixel image correspondences is a fundamental task of computer vision. Often, the objective is to align image pairs from the same semantic category for manipulation or segmentation purposes. Despite achieving superior performance, existing deep learning alignment methods cannot cluster images; consequently, clustering and pairing images needed to be a separate laborious and expensive step. Given a dataset with diverse semantic categories, we propose a multi-task model, Jim-Net, that can directly learn to cluster and align images without any pixel-level or image-level annotations. We design a pair-matching alignment unsupervised training algorithm that selectively matches and aligns image pairs from the clustering branch. Our unsupervised Jim-Net achieves comparable accuracy with state-of-the-art supervised methods on benchmark 2D image alignment dataset PF-PASCAL. Specifically, we apply Jim-Net to cryo-electron tomography, a revolutionary 3D microscopy imaging technique of native subcellular structures. After extensive evaluation on seven datasets, we demonstrate that Jim-Net enables systematic discovery and recovery of representative macromolecular structures in situ, which is essential for revealing molecular mechanisms underlying cellular functions. To our knowledge, Jim-Net is the first end-to-end model that can simultaneously align and cluster images, which significantly improves the performance as compared to performing each task alone.

The growth of cutaneous T-cell lymphoma is stimulated by immature dendritic cells.

In the initial stage of cutaneous T-cell lymphoma (CTCL), proliferating CTCL cells are concentrated in the epidermis in close association with an immature dendritic cell (DC), the Langerhans cell. Because long-term in vitro culture of CTCL cells has proven difficult, the in vivo association with the major antigen-presenting cell (APC) of the epidermis has been postulated to play a role in directly stimulating the clonal T-cell proliferation. We report that CTCL cells can be reproducibly grown in culture for 3 months when cocultured with immature DCs. CTCL cells retain the phenotype and genotype of the initial malignant clone, whereas the APCs are a mixture of immature and mature DCs. CTCL cell and DC survival was dependent on direct membrane contact. Growth was inhibited by antibodies that bound to the T-cell receptor (TCR) or interfered with the interaction of CD40 with its ligand on the CTCL cell. Addition of antibody to CD3 or the clonotypic TCR caused rapid CTCL cell apoptosis followed by engulfment by avidly phagocytic immature DCs and subsequent DC maturation. The opportunity to study CTCL cells and immature DCs for prolonged periods will facilitate studies of tumor cell biology and will allow investigation of the intriguing hypothesis that CTCL cell growth is driven through TCR recognition of class II-presented self-peptides. In addition, the culture of CTCL cells will permit evaluation of therapies in vitro before clinical intervention, thereby improving safety and efficacy.