top of page
Metabolism-banner-01.png

Metabolic Imaging

AI-driven imaging of cellular functions and metabolic networks

We develop techniques based on the “metabolic contrast”. While in traditional imaging probes are used to visualize morphological structures exploiting differences in qualities such as density or water content, in metabolic imaging we are able to image energy trasformations, by making probes interact chemically with their surroundings and in turn alter the image according to molecular changes occurring within the area of interest.  For example, we developed some methods to image redox potential of glutathione in cells with sub-nanometric resolution, as well as pH and membrane fluidity. 

Learning glucose uptake dynamics 
1-s2.0-S0010482522002153-gr3 (1).jpg

Automated detection and classification of tumor histotypes on dynamic PET imaging data through machine-learning driven voxel classification

Computers in Biology

and Medicine

June 1, 2022

2-deoxy-2-fluorine-(18F)fluoro-d-glucose Positron Emission Tomography/Computed Tomography (18F-FDG-PET/CT) is widely used in oncology mainly for diagnosis and staging of various cancer types, including lung cancer, which is the most common cancer worldwide. Since histopathologic subtypes of lung cancer show different degree of 18F-FDG uptake, to date there are some diagnostic limits and uncertainties, hindering an 18F-FDG-PET-driven classification of histologic subtypes of lung cancers. On the other hand, since activated macrophages, neutrophils, fibroblasts and granulation tissues also show an increased 18F-FDG activity, infectious and/or inflammatory processes and post-surgical and post-radiation changes may cause false-positive results, especially for lymph-nodes assessment. Here we propose a model-free, machine-learning based algorithm for the automated classification of adenocarcinoma, the most common type of lung cancer, and other types of tumors. Input for the algorithm are dynamic acquisitions of PET data (dPET), providing for a spatially and temporally resolved characterization of the uptake kinetic. The algorithm consists in a trained Random Forest classifier which, relying contextually on several spatial and temporal features of 18F-FDG uptake, generates as an outcome probability maps allowing to distinguish adenocarcinoma from other lung histotype and to identify metastatic lymph-nodes, ultimately increasing the specificity of the technique. Its performance, evaluated on a dPET dataset of 19 patients affected by primary lung cancer, provides a probability 0.943 ± 0.090 for the detection of adenocarcinoma. The use of this algorithm will guarantee an automatic and more accurate localization and discrimination of tumors, also providing a powerful tool for detecting at which extent tumor has spread beyond a primary tumor into lymphatic system.

Red Blood Cells are biosensors
1-s2.0-S259013461930026X-fx1.jpg

Erythrocyte membrane fluidity as a marker of diabetic retinopathy in type 1 diabetes mellitus

European Journal of Clinical

Investigation

May 1, 2021

A high level of glycosylated haemoglobin (HbA1c), which is a nonenzymatic glycosylation product, is correlated with an increased risk of developing microangiopathic complications in Diabetes Mellitus (DM). Erythrocyte membrane fluidity could provide a complementary index to monitor the development of complications since it is influenced by several hyperglycaemia-induced pathways and other independent risk factors.

15 healthy controls and 33 patients with long-duration (≥20 years) type 1 Diabetes Mellitus (T1DM) were recruited. Diabetic subjects were classified into two groups: T1DM, constituted by 14 nonretinopathic patients, and T1DM + RD, constituted by 19 patients in any stage of diabetic retinopathy. Red blood cells (RBC) were incubated with the fluorescent Laurdan probe and median values of Generalized Polarization (GP), representative of membrane fluidity, were compared between the two groups. Baseline characteristics among groups have been compared with Student's t test or ANOVA. Values of P < .05 were considered statistically significant.

All the participants were comparable for age, Body Mass Index (BMI), creatinine and lipid profile. The duration of diabetes was similar for T1DM (34.4 ± 7.8 years) and T1DM + RD (32.8 ± 7.5 years) subjects as well as values of HbA1c: (55.6 ± 8.1) mmol/mol for T1DM and (61.2 ± 11.0) mmol/mol for T1DM + RD, respectively. Erythrocyte plasmatic membranes of RD patients were found to be more fluid (GP: 0.40 ± 0.04) than non-RD patients (GP: 0.43 ± 0.03) with a statistically significant difference (P = .035).

Altered erythrocyte membrane fluidity may therefore represent a marker of retinopathy in T1DM patients as a result of post-translational modifications of multifactorial aetiology (nonenzymatic glycosylation of proteins, generation of reactive oxygen species, lipid peroxidation).

Label Free Artificial Intelligence
1-s2.0-S000326702031254X-fx1_lrg.jpg

Label-free metabolic clustering through unsupervised pixel classification of multiparametric fluorescent images

Analytica Chimica Acta

March 1, 2021

Autofluorescence microscopy is a promising label-free approach to characterize NADH and FAD metabolites in live cells, with potential applications in clinical practice. Although spectrally resolved lifetime imaging techniques can acquire multiparametric information about the biophysical and biochemical state of the metabolites, these data are evaluated at the whole-cell level, thus providing only limited insights in the activation of metabolic networks at the microscale. To overcome this issue, here we introduce an artificial intelligence-based analysis that, leveraging the multiparametric content of spectrally resolved lifetime images, allows to detect and classify, through an unsupervised learning approach, metabolic clusters, which are regions having almost uniform metabolic properties. This method contextually detects the cellular mitochondrial turnover and the metabolic activation state of intracellular compartments at the pixel level, described by two functions: the cytosolic activation state (CAF) and the mitochondrial activation state (MAF). This method was applied to investigate metabolic changes elicited in the breast cancer cell line MCF-7 by specific inhibitors of glycolysis and electron transport chain, and by the deregulation of a specific mitochondrial enzyme (ACO2) leading to defective aerobic metabolism associated with tumor growth. In this model, mitochondrial fraction undergoes to a 13% increase upon ACO2 overexpression and the MAF function changes abruptly by altering the metabolic state of about the 25% of the mitochondrial pixels.

Artificial Intelligence in search of new phases
boe-11-10-5728-g001.jpg

Unsupervised clustering of multiparametric fluorescent images extends the spectrum of detectable cell membrane phases with sub-micrometric resolution

Biomedical Optics Express

October 1 , 2020

Solvatochromic probes undergo an emission shift when the hydration level of the membrane environment increases and are commonly used to distinguish between solid-ordered and liquid-disordered phases in artificial membrane bilayers. This emission shift is currently limited in unraveling the broad spectrum of membrane phases of natural cell membranes and their spatial organization. Spectrally resolved fluorescence lifetime imaging can provide pixel-resolved multiparametric information about the biophysical state of the membranes, like membrane hydration, microviscosity and the partition coefficient of the probe. Here, we introduce a clustering based analysis that, leveraging the multiparametric content of spectrally resolved lifetime images, allows us to classify through an unsupervised learning approach multiple membrane phases with sub-micrometric resolution. This method extends the spectrum of detectable membrane phases allowing to dissect and characterize up to six different phases, and to study real-time phase transitions in cultured cells and tissues undergoing different treatments. We applied this method to investigate membrane remodeling induced by high glucose on PC-12 neuronal cells, associated with the development of diabetic neuropathy. Due to its wide applicability, this method provides a new paradigm in the analysis of environmentally sensitive fluorescent probes.

Tracking lipid metabolism
Senzanome.gif

Real time quantitative analysis of lipid storage and lipolysis pathways by confocal spectral imaging of intracellular micropolarity

Biochimica Biophysica Acta

Molecular and Cell Biology of Lipids

April 11, 2018

Organisms store fatty acids in triacylglycerols in the form of lipid droplets, or hydrolyze triacylglycerols in response to energetic demands via activation of lipolytic or storage pathways. These pathways are complex sets of sequential reactions that are finely regulated in different cell types. Here we present a high spatial and temporal resolution-based method for the quantification of the turnover of fatty acids into triglycerides in live cells without introducing sample preparation artifacts.

We performed confocal spectral imaging of intracellular micropolarity in cultured insulin secreting beta cells to detect micropolarity variations as they occur in time and at different pixels of microscope images. Acquired data are then analyzed in the framework of the spectral phasors technique.

The method furnishes a metabolic parameter, which quantitatively assesses fatty acids - triacylglycerols turnover and the activation of lipolysis and storage pathways. Moreover, it provides a polarity profile, which represents the contribution of hyperpolar, polar and non-polar classes of lipids. These three different classes can be visualized on the image at a submicrometer resolution, revealing the spatial localization of lipids in cells under physiological and pathological settings.

This new method allows for a fine-tuned, real-time visualization of the turnover of fatty acids into triglycerides in live cells with submicrometric resolution. It also detects imbalances between lipid storage and usage, which may lead to metabolic disorders within living cells and organisms.

Autophagic flux unveiled
kaup-11-10-1084455-g001.jpg

Quantitative analysis of autophagic flux by confocal pH-imaging of autophagic intermediates

Autophagy

October 11, 2015

Although numerous techniques have been developed to monitor autophagy and to probe its cellular functions, these methods cannot evaluate in sufficient detail the autophagy process, and suffer limitations from complex experimental setups and/or systematic errors. Here we developed a method to image, contextually, the number and pH of autophagic intermediates by using the probe mRFP-GFP-LC3B as a ratiometric pH sensor. This information is expressed functionally by AIPD, the pH distribution of the number of autophagic intermediates per cell. AIPD analysis reveals how intermediates are characterized by a continuous pH distribution, in the range 4.5–6.5, and therefore can be described by a more complex set of states rather than the usual biphasic one (autophagosomes and autolysosomes). AIPD shape and amplitude are sensitive to alterations in the autophagy pathway induced by drugs or environmental states, and allow a quantitative estimation of autophagic flux by retrieving the concentrations of autophagic intermediates.

bottom of page
Privacy Policy Terms and Conditions