npj Imaging最新文献

Glutamine metabolism is upregulated in many cancers. While multiple glutamine imaging agents have been developed and translated to clinical use, the short half-lives of their signal and instability in vivo limit the aspects of glutamine metabolism they capture. In phantoms at physiological pH, chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) contrast was observed at 11.7 T from glutamine, downstream metabolic products (glutamate and ammonia) and their co-substrates (alanine, aspartate, and cystine/cysteine). This contrast increased at lower pH. These results suggest that both uptake and metabolism of glutamine would increase CEST signal enhancement. We then investigated the feasibility of imaging the uptake (delivery, transport and metabolism) of naturally-occuring glutamine using CEST MRI in preclinical prostate cancer models, wherein key metabolic proteins are the glutamine transporter ASCT2 and as well as enzymes GLS1, ALT2 (GPT2), AST1 (GOT1), and GDH1 (GLUD1). The LNCaP prostate cancer line exhibited higher expression of ASCT2, GDH1, ALT2, and AST1 compared to DU-145 cells. CEST MRI enhancement upon administration of glutamine was consistently higher in LNCaP 3D spheres (phantoms) and tumors (in vivo) than their DU-145 counterparts. Mass spectrometry imaging confirmed higher uptake and metabolism of glutamine in LNCaP tumors. These findings demonstrate that CEST MRI of glutamine is capable of distinguishing preclinical prostate tumor models that differ in glutamine uptake and has potential for translation to clinical use.

Mutations in mitochondrial-related genes underlie numerous neurodegenerative diseases, yet the significance of most variants remains uncertain concerning disease phenotypes. Several thousand genes have been shown to regulate mitochondria in eukaryotic cells, but which of these genes are necessary for proper mitochondrial function and dynamics? We investigated the degree of morphological disruptions in mitochondrial gene-silenced cells to understand the genetic contribution to the expected mitochondrial phenotype and to identify potentially pathogenic variants like pathogenic mutations in MFN2. We analyzed 5835 gRNAs in a high dimensional phenotypic dataset produced by the image-based pooled analysis platform Raft-Seq. Using the MFN2-mutant cell phenotype, we identified several genes, including TMEM11, TIMM8A, NDUFAF4, NDUFAF7, and NDUFS5 (NADH ubiquinone oxidoreductase-related genes), as crucial for normal mitochondrial dynamics in human U2OS cells. Additionally, we found several missense and UTR variants within the genes SLC25A19 and ATAD3A as drivers of mitochondrial aggregation. By examining multiple features instead of a single readout, this analysis was powered to detect genes which had morphological 'signatures' aligned with MFN2-mutant phenotypes. Reanalysis with anomaly detection revealed other critical genes, including APOOL, MCEE, NIT, PHB, and SLC16A7, which perturb mitochondrial network morphology in a manner divergent from MFN2. These studies show causal links between gene knockouts and gene-specific variants into the assembly or maintenance of mitochondrial dynamics and can hopefully lead to a better understanding of mitochondrial related diseases.

Axonal myelination finely tunes action potential conduction to control precise timing in neural circuits. Little is known about how dynamic myelinating oligodendrocytes are in the adult brain primarily due to limited approaches for their investigation at the cellular level over time in their native environment. This protocol describes optical imaging approaches that allow specific label-free detection of compact myelin which, when combined with genetically encoded fluorescence reporters and small molecule dyes, permits high-resolution longitudinal and fixed sample imaging of myelin and oligodendrocytes in live mice, in live organotypic slices, and in postmortem tissues. Data generated with these approaches can be used to test fundamental questions related to myelin development, plasticity, maintenance, and repair.

Analysis of biological images relies heavily on segmenting the biological objects of interest in the image before performing quantitative analysis. Deep learning (DL) is ubiquitous in such segmentation tasks, but can be cumbersome to apply, as it often requires a large amount of manual labeling to produce ground-truth data, and expert knowledge to train the models. More recently, large foundation models, such as SAM, have shown promising results on scientific images. They, however, require manual prompting for each object or tedious post-processing to selectively segment these objects. Here, we present FeatureForest, a method that leverages the feature embeddings of large foundation models to train a random forest classifier, thereby providing users with a rapid way of semantically segmenting complex images using only a few labeling strokes. We demonstrate the improvement in performance over a variety of datasets and provide an open-source implementation in napari that can be extended to new models.

We demonstrate that rigid anchoring of fluorescent proteins through double tagging (FPs) in living cells can significantly enhance the contrast in fluorescence polarization microscopy (FPM) by locking the transition dipole moment orientations to the sample's structures. We applied double tagging of reversibly photoswitchable FPs (dt-rsFPs) to membranes and present a novel camera frame-separated switching pulse scheme that allows effective narrowing of the angle range of excited dt-FP also in living cells (frame-separated excitation polarization angle narrowing, FrExPAN). The principle of rigid anchoring allows specific selection of signals from different structural cell parts with slightly different orientations and is broadly applicable. FrExPAN imaging with dt-rsFPs double-tagged to membranes of living HeLa cells and living hippocampal neurons is demonstrated. We discuss potential implications for orientational contrast imaging as well as super-resolution by polarization demodulation (SPoD) methods.

Mild traumatic brain injury (mTBI) is neurological impairment induced by biomechanical forces without structural brain damage, currently without an objective diagnostic tool. Downstream injury stems from oxidative damage leading to the production of neurotoxic aldehydes. A collagen-based 3D corticomimetic in vitro model of concussion was developed, confirming aldehyde production following impact. Total aldehyde levels were mapped in vivo following mTBI using a novel CEST-MRI contrast agent, ProxyNA₃, in a new model of closed-head, awake, single-impact concussion in aged and young mice with aldehyde dehydrogenase 2 (ALDH2) deficiency. ProxyNA₃-MRI was performed before impact, and on days two- and seven- post-impact. MRI signal enhancement significantly increased at two days post-injury prior to astrocyte activation at seven days post-injury. The data suggest that advanced age and ALDH2 deficiency contribute to increased aldehydic load following mTBI. Overall, ProxyNA₃ was capable of mapping concussion-associated aldehydes, supporting its application as an objective diagnostic tool for concussion.

Metabolic reprogramming is considered a major driving factor in cancer growth and yet it remains challenging to monitor in vivo uptake of fatty acids, which are essential energy sources for many tumor types. Here, we report the development of a novel, long-chain fatty acid (FA), near-infrared (NIR) imaging reagent (FA-ICG) for real-time, non-invasive imaging of FA absorption in vitro and in vivo. Moreover, we demonstrate the application of the probe in image-guided cancer surgery, where precise assessment of tumor margins is paramount for removal. Specifically, we focus on glioblastoma (GBM), where FA metabolism plays a key role in progression and where there is a significant need for better intraoperative imaging. Here, we successfully demonstrate the application of the probe for NIR in vivo imaging in two different orthotopic models of GBM. In addition, we validate the uptake of the probe in companion dogs with mastocytomas, as these develop cancer with a similar pathology to humans. Our results demonstrate that the probe combines benefits from NIR imaging, such as high sensitivity, low autofluorescence, and deep tissue penetration, with specific tumor metabolism-based targeting and retention. Thus, it represents a promising candidate for a wide range of applications in the fields of metabolic imaging, drug development, and most notably for translation in image-guided surgery.

Functional magnetic resonance imaging (fMRI), exploiting the blood oxygen level-dependent (BOLD) contrast, is the most widely used technique to study brain function. Combined with tools from biotechnology, molecular biology, and genetics, preclinical fMRI offers unparalleled opportunities to experimentally test causal hypotheses that are beyond the reach of human research. Here, we review recent progress in MRI hardware development, provide recommendations for BOLD fMRI protocol optimization, and discuss recent applications.