Portal de Periódicos da CAPES

Google maps for tissues: Multiscale imaging of biological systems and disease

2019; Wiley; Volume: 228; Issue: 2 Linguagem: Inglês

10.1111/apha.13392

1748-1716

Thoralf Niendorf, Lucio Frydman, Michal Neeman, Erdmann Seeliger,

Medical Imaging Techniques and Applications

Seeing is believing. The physicist Richard Feynman highlighted in his landmark lecture in 1959 "… that it is very easy to answer many of these fundamental biological questions; you just look at the thing!" Considering the spatial resolution constraints of imaging and foreseeing the challenges of big data and quantification Feynman concluded: "You should use more mathematics. Make the microscope one hundred times more powerful, and many problems of biology would be made very much easier." Since then, many disruptive technologies and new directions in life science and medicine have been launched - driven often by the emergence of new biomedical imaging tools. Evidence of this is clear from the 2003 Nobel Prize for magnetic resonance imaging (MRI), the 2014 Nobel Prize for advanced light microscopy, and the 2017 Nobel Prize for developments in cryo-electron microscopy. Progress in breaking new spatial and temporal boundaries, hybrid imaging modalities, hardware developments boosting sensitivity and lowering detection levels, artificial intelligence and deep machine learning, sophisticated image fusion approaches and explorations into big data science are some of the elements propelling the ongoing revolution in biomedical imaging. Additional driving forces come from advances in photonics and algorithmics that have resulted in super-resolved optical imaging techniques. Advances in genetics and genomics have spurred developments and applications in the optical imaging of cellular activity, including the visualization of gene expression in real-time. New imaging tools including advanced hyperpolarized MRI and ultrahigh field MRI that provide new contrasts and sensitivity gains have had disruptive effects, and will surely become indispensable components at the interface of physics, data sciences, physiology and clinical care. One of the most exciting areas of innovation in biomedical research concerns the visualization on multiple scales in space and time: imaging biological objects in size ranging from Ångströms and nanometers to the scale of whole body imaging, and from picoseconds to decades in large population imaging studies. In the Century of Biology, enabled by the tools of information technology and artificial intelligence, one can imagine a concept like 'Google Maps' for each fundamental biological research question, in which views at the macroscopic, mesoscopic, microscopic and the nanoscopic scales are stitched together for seamless zooming, depending on the needs and interests of the scientists and users, the research fields and the patients. Considering the imaging pipeline, the push for a google maps for tissues approach involves not only the advancement of image acquisition and image reconstruction techniques but also innovations in computational sciences for the information integration of data obtained across multiple imaging modalities and scales in space and time with the goal to be able to overlay large datasets from multiple measurements into one coherent representation that can be used for visualization, annotating and processing. Answering these questions is an area of vigorous ongoing research, and only few valuable recent developments and applications will be exemplarily highlighted here. On the mesoscopic or more macroscopic scale, the establishment of advanced in vivo anatomic and functional phenotyping, non-invasive tissue characterization as an alternative to invasive biopsies, and in vivo histology workflows all facilitated by MRI are conceptually appealing for the pursuit of imaging in support of translational research as demonstrated by outstanding examples published in Acta Physiologica. The capabilities of parametric MRI using the MR relaxation parameter T1 for non-invasive tissue characterization were employed to demonstrate that overexpression of integrin α11 induces cardiac fibrosis and left ventricular hypertrophy in mice.1 Another group of MR techniques for tissue characterization makes use of the relaxation time T2 to probe for tissue water changes or the apparent water diffusion coefficient (ADC) to monitor changes in the intra- and extracellular volume. These MRI metrics are instrumental to monitor cerebrovascular changes: a recent study revealed that acute mitogen-activated protein kinase 1/2 inhibition improves functional recovery and vascular changes after ischaemic stroke by means of in vivo T2 and ADC mapping in rats.2 Dynamic contrast enhanced (DCE)-MRI can be used for characterization of tissue perfusion. Applied to the kidney, it provides information about renal perfusion and glomerular filtration rate (GFR). This MR renography approach was recently applied to study the response to stimulation of olfactory calcium-sensing receptors (CaSR) and to detail the pathways and mechanisms of this response in rats.3 Increased renal sympathetic nervous activity depressed GFR within 30 seconds upon stimulation of intranasal CaSR.3 Unlike MR renography, the measurement of GFR by clearance techniques would require much longer sampling periods. This is prohibitive for the study of very short-term effects such as olfactory CaSR activation and highlights the benefits and value of DCE-MRI based characterization. Other highlights of DCE-MRI applications include monitoring of microcirculation to evaluate the role of exchange protein activated by cyclic adenosine monophosphate 1 in the response of transvascular fluid exchange to histamine in mice.4 Parametric MRI enables in vivo temperature mapping that opens an entirely new imaging driven research field of thermal phenotyping. The relevance of MR thermometry for the study of metabolically active brown adipose tissue in humans inspired a debate which is formed around the question whether skin temperature may not yield human brown adipose tissue activity in diverse populations because of the shortcomings and physiological confounders of infrared based thermometry.5 On the mesoscopic scale emerging photoacoustic and ultrasound (US) imaging present an alternative to MRI. Recent work offered a fascinating depiction of how US approaches can extend the spectrum of clinical applications into the domain of basic and applied physiological research. A striking example from human integrative muscle physiology revealed that a reduction in activation level reduces residual force depression in tibialis anterior muscle.6 For the assessment of muscle architecture changes including fascicle length and pennation angle, B-mode US was employed using a temporal resolution of 16 ms (framerate: 60 Hz).6 Trevino et al used US to probe muscle cross-sectional area and muscle tissue composition (contractile vs non-contractile tissue), thereby revealing details of sex-related differences in trapezoid muscles.7 On the microscopic scale optical imaging highlights recently published in Acta Physiologica include the application of transmission electron microscopy for the assessment of mitochondrial densities and profiles in humans.8 This approach revealed that exercise training increases skeletal muscle mitochondrial volume density by enlargement of existing mitochondria and not de novo biogenesis.8 Winje et al used microscopic imaging to address the question if cachexia—a severe wasting disorder including muscle atrophy—is a mere perturbation of the protein balance or if the condition also involves a degenerative loss of myonuclei within the fibre syncytia or loss of whole muscle fibres.9 Single live fibres in the extensor digitorum longus muscle of mice were injected a solution containing oligonucleotides that serves as an intravital nuclear dye for counting myonuclei.9 Utilizing in vivo microscopy of single muscle fibres, as well as conventional immunohistochemical ex vivo analysis, the authors detected no significant fibre loss and no reduction in myonuclear number in spite of a 21% reduction in fibre size.9 It was concluded that with no loss of cells or myonuclei, targets for therapeutic interventions could be narrowed down to intracellular signalling pathways of the atrophying muscle fibres.9 In vivo Ca2+ imaging is a powerful tool in physiology, which provides temporal resolution at the millisecond range and spatial resolution at micrometer range. Recent advancements were enabled by novel designs of fluorescent Ca2+ sensors, development of modern microscopes and powerful imaging techniques such as two-photon microscopy. Applications range from imaging Ca2+ sparks in cardiomyocytes to imaging of neurones and neuroendocrine cells.10 A recent study took advantage of mechanical stretch-induced real time Ca2+ imaging in isolated esophageal myenteric neurones (EMN) to demonstrate that the Na+/Ca2+ exchanger 1 plays an important role for the mechanosensitivity in the EMN of rats and humans.11 In another report Ca2+ imaging was employed to investigate morphologically distinct paraneurones of the urethra - a class of specialized cells that express serotonin—and their potential role in peripheral sensory information processing.12 On the nanoscopic scale, development and application of next generation optical microscopy encompasses single molecule analyses and super-resolution microscopy in signalling research. This is of high relevance since receptors and their signals are among the main biological processes that are exploited for drug therapy. Using labelling strategies with organic fluorophores, they can now be studied in intact cells at the single molecule level, with a spatial resolution down to ≈30 nm and a temporal resolution of about 30 ms. Integration of single cell microscopy and single cell genomics is another pioneering area of activity. In current biomedical imaging, most decisions are made on the basis of qualitative markers or surrogates. This approach renders interpretation of imaging results highly subjective, requires thousands of hours of training and limits if not prohibits reproducibility of data across observers, sites, modalities and vendors. Every image is based on the measurement of physical quantities though. Imaging a metric facilitates objective assessment and quantification of data. This includes the quantification and mapping of parameters related to biophysics and physiology over a wide range of spatial and temporal scales. These research efforts are committed to making biomedical imaging a more quantitative science, which is essential to detail the links between imaging markers, molecular profiles, biochemical markers and the signatures from physiological measurements and to decipher what biomedical imaging signatures tell us about biological function and evolution of diseases. En route to imaging parameters that are clinically usable quantitative biomarkers, calibration by means of gold standard quantitative methods is an essential step and spring board. Ideally, the imaging and the quantitative measurements - often by techniques established in laboratories of integrative physiology - should be done simultaneously in the same individual. As all established modalities available in today's experimental and translational research, these gold standard techniques have shortcomings and limitations. A particular limitation of many physiological methods is their invasiveness, thus, the calibration must, in general, be done by pre-clinical studies. Functional MRI of the kidney may serve as a striking example: state-of-the-art renal MRI allows investigators to qualitatively assess physiological parameters such as renal blood flow, GFR and oxygenation.13, 14 Renal MRI is non-invasive, affords full kidney coverage, excellent soft tissue contrast, high temporal resolution, and longitudinal studies without ionizing radiation,15, 16 as illustrated by Figure 1 showing multi-parametric MRI for probing different physiological parameters at different spatial scales using T1, T2, T2*, blood oxygenation level dependent, apparent water diffusion and renal blood volume contrast. Yet, the validity of functional MRI techniques for renal parameters for various (patho-)physiological scenarios remains to be established.15 Quantitative data on functional renal parameters obtained by MRI, without proper calibration, may be erroneous and their interpretation is questionable.17, 18 To address this issue parametric MRI needs to be calibrated with quantitative gold standard methods. With regard to renal oxygenation, calibration is on its way: By means of an integrated multi-modality approach designated as MR-PHYSIOL the simultaneous tracking of invasive physiological parameters and the oxygenation-sensitized MR parameter T2* was demonstrated in the same kidney of rats in vivo.15, 19 The pace of discovery in biomedical imaging is heartening, drawing in new talent and driving transfer of the results into preclinical applications and the clinical arena. The remaining challenges should be faced openly in collaborations between forward-thinking researchers, application scientists and clinicians. They should be interdisciplinary, inter-institutional and international, as exemplified and spearheaded by imaging networks (for example: http://www.eurobioimaging.eu/). From today's base camp, complex and cost intense imaging technology may look like mountainous heights; inevitably, these peaks will be the easily accessible green valley meadows of tomorrow. With this momentum imaging has become an integral part of large scale epidemiological studies (for example: UK biobank, https://www.ukbiobank.ac.uk/; German National Cohort, NAKO, https://nako.de/); that provide a comprehensive and unprecedented deep phenotyping and biosampling with the ultimate goal to understand the inherited and acquired determinants of health in populations and to shape the future of personalized prevention.20 Sharing complementary expertise, state-of-the-art infrastructure and outstanding research expertise through Imaging Centers and distributed research infrastructure (for example: ELIXIR, https://elixir-europe.org/) promotes access to and boosts usage of imaging technology and substantially enhances the scientific contributions of biomedical imaging. To take this approach to the next level, calls for National Research Infrastructure initiatives have been launched (for example: https://www.bmbf.de/de/roadmap-fuer-forschungsinfrastrukturen-541.html) driving the establishment of well-funded National Imaging Facilities (for example: Australian National Imaging Facility, https://anif.org.au/). And there will be the inevitable breakthroughs and surprises that come when you place next-generation imaging instruments in the hands of highly creative interdisciplinary teams. However, this will only happen if we recognize that moving into next generation of biomedical imaging technology is more than just a matter of buying equipment and installing them and trying to operate them in "core facilities" where budgetary and not scientific goals and considerations might be dominant. The ultimate potential of this game-changing technology is far greater; all that is required is the imagination to apply it. The authors acknowledge the members of the Berlin Ultrahigh Field Facility (B.U.F.F.) at the Max-Delbrueck Center for Molecular Medicine in the Helmholtz Association, Berlin, Germany; the faculty and supporters of the Imaging from the NAno to the Meso (iNAMES) consortium and the colleagues and collaborators from the Institute of Physiology, Charité – Universitätsmedizin Berlin Germany; who kindly contributed examples and summaries of their pioneering work or other valuable assistance. The authors have no conflict of interest to declare. This work was funded in part (T. Niendorf, E. Seeliger) by the German Research Foundation (Gefördert durch die Deutsche Forschungsgemeinschaft (DFG)—Projektnummer 394046635—SFB 1365. Funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation)—Projekt nummer 394046635—SFB 1365), by the DZHK (T. Niendorf, German Centre for Cardiovascular Research, BER 6.1, partner site Berlin) and by the Federal Ministry of Education and Research, Berlin, Germany, FKZ 81Z6100161.