Super-Resolution Microscopy Writ Large

news and views

Super-resolution microscopy writ large

Peter Engerer, Caroline Fecher & Thomas Misgeld

Improved expansion microscopy provides a super-resolution method that is accessible to any laboratory with a fluorescence microscope.

Seeing is believing, but many subcellular struc- a tures are simply too small and densely packed to be resolved with light microscopes. Over the past two decades, ever-more sophisticated techniques for ‘super-resolution’ microscopy have succeeded in surpassing the capabilities of conventional microscopy. And last year, a new technology called expansion microscopy1 achieved the precision of super-resolution imaging not by optical tricks but by a radically different approach in which the biological specimen is physically enlarged through chemical treatments. In this issue and in a recent issue b of Nature Methods, three groups, Tillberg et al.2, Ku et al.3, and Chozinski et al.4, improve on this method with simplified protocols that use off-the-shelf chemicals and are compatible with standard immunofluorescence techniques and genetically encoded fluorophores. The new protocols should broaden the applications of Nature America, Inc. All rights reserved. America, Inc. © 201 6 Nature expansion microscopy and bring it within reach of almost any laboratory. Light microscopy has contributed tremen- npg dously to our understanding of cell biology. But the physics of light diffraction limits the spatial resolution of conventional microscopes to ~200 nm in the lateral dimension and ~500 nm in the axial dimension, which means that distinct structures in closer proximity cannot be resolved. Many subcellular organelles are smaller than 200 nm and are densely packed, eluding visualization as individual structures by light microscopes. For many years, only electron microscopy could circumnavigate the diffraction limit of light, but conventional

Peter Engerer, Caroline Fecher and Thomas Misgeld are at the Institute of Neuronal Cell Biology, Munich, Germany, the Center of Integrated Protein Science (CIPSM), Munich, Germany, the German Center for Neurodegenerative Diseases (DZNE), Munich, Figure 1 Concept and workflow of refined expansion microscopy techniques. (a) Expansion microscopy Germany, and the Munich Cluster of Systems expands tissue embedded in a hydrogel, similar to how the expansion of a balloon reveals details of a Neurology (SyNergy), Munich, Germany. sketch drawn on its surface (i.e., in 2D). (b) Simplified comparative workflow of different expansion e-mail: [email protected] microscopy protocols.

928 volume 34 number 9 September 2016 nature biotechnology news and views

Table 1| Comparison of demonstrated expansion microscopy techniques as reported ExM1 ExM4 proExM2 MAP3 ExFISH9 Boyden laboratory Vaughan laboratory Target Proteins Proteins and DNA (dye) Proteins Proteins and sugar RNA and DNA (dye) residues Matrix 2.5% acrylamide 2.5% acrylamide sodium 2.5% acrylamide sodium 30% acrylamide, 2.5% acrylamide sodium sodium acrylate and acrylate and MBAA acrylate and MBAA sodium acrylate acrylate and MBAA MBAA and BA Linking agent Trifunctional DNA- MA-NHS, AcX None LabelX, (AcX plus Label-IT oligomer GA amine) Sample disruption Digest proteinase K, Digest proteinase K, 30 min Digest proteinase K, >12 Denaturation with Digest proteinase K, >12 >12 h, RT to >12 h, 37 °C h, RT; 4 h, 60 °C SDS, 37 °C and dis- h, 37 °C sociation at 70/95 °C, 6 h Expansion factor 4.5 4.0–4.2 ~4.0 4.0 3.3 (resolution) (~70 nm) (65 nm) (~70 nm) (~60 nm) (not reported) Time ~6 d ~4.5 d ~3.0 d ~7 d ~4.5 d (brain slice) (100 µm) (100 µm) (100 µm) (100–500 µm) (50–200 µm) FP preservation No Yes, max. digest 30 min Yes, ~50% intensity No Not applicable (combina- to 1 h tion with proExM2) IF staining No Yes, Yes Yes No modified mAb Sample Cells, tissue: brain Cells, tissue: brain Cells, tissue: brain, pan- Cells, tissue: brain, Cells, tissue: brain creas, lung, spleen lung, heart, spinal cord, liver, kidney, intestine

Comment First report of the Pre-expansion staining Pre-expansion staining; Post-expansion Post-expansion FISH with concept, customized post-expansion staining staining of preserved multiplexing and HCR reagents needed possible epitopes; multiplexing amplification AcX, Acryloyle-X; mAb, antibody; BA, bisacrylamide; FISH, fluorescencein situ hybridization; FP, fluorescent protein; GA, glutaraldehyde; HCR, hybridization chain reaction; MA-NHS, methacrylic acid-N-hydroxysuccinimidyl ester; MBAA, N,N’-methylenebisacrylamide; RT, room temperature.

electron microscopy is extremely laborious of specimens is inadvertently altered6. It takes Now, the intersection of repetitive immunos- and restricted to minuscule tissue volumes, advantage of this effect by increasing the extent taining and tissue expansion makes it possible and offers only limited options for labeling and isotropy of the expansion to enlarge the to study at high resolution a wide array of anti- specific antigens or other molecular targets. specimen about fourfold and achieve an effec- gens (although, currently, nothing close to a full More recently, super-resolution fluorescence tive lateral resolution of ~70 nm on conventional proteome)3. MAP achieves expansion through Nature America, Inc. All rights reserved. America, Inc. © 201 6 Nature microscopy has improved resolution of light light microscopes1. First-generation expansion preventing intra- and interprotein crosslinking microscopes by about an order of magnitude microscopy was carried out by labeling pro- during the hydrogel–sample hybridization. By (in optimal cases as low as ~20 nm lateral tein antigens, embedding the sample in a applying a high concentration of acrylamide npg resolution) through techniques for spatially hydrogel, tethering the label to the hydrogel, the interaction of sample and hydrogel is maxi- shaping the illumination and exploiting sto- digesting the original tissue, and expanding the mized and isotropic expansion is achieved after chastic variations in fluorescence emission5. hydrogel ‘blue print’ (Fig. 1b). But because any high-temperature denaturation without using However, super-resolution techniques are fluorescent protein or antibody used for label- proteinase K. not in wide use owing to the high cost of the ing would be degraded, it was necessary to use Expansion microscopy has already been microscope, the need for specialized expertise, specialized, non-protein reagents. applied successfully to several tissues and spe- or other inherent shortcomings (e.g., the dif- The new studies2–4 address this limitation in cies, and it is likely to be adaptable to a wide ficulty of multicolor experiments, the limited different ways (Fig. 1b and Table 1). Tillberg range of specimens. As these applications are imaging depth, the high power of excitation et al.2 and Chozinski et al.4 introduce expansion developed, the isotropy of the expansion will light, and the slow acquisition speed). microscopy variants, which preserve proteins remain a key concern, especially for tissues Expansion microscopy promises to over- in the sample during the expansion process to a with special mechanical properties, such as come a number of these practical and scientific degree that allows standard fluorescent protein muscles and tendons. Thus, similar to previous limitations. In essence, the approach amounts fusions and off-the-shelf secondary antibodies new fixation and microscopy techniques, it will to taking an organ, fixing it, and blowing it up to remain detectable. This is achieved by be important to validate results and distinguish like a balloon (Fig. 1a). Could it really be this anchoring proteins to the hydrogel and facts from artifacts by painstaking comparative simple to bypass the resolution limit of light minimizing proteolytic homogenization by analysis across different imaging modalities. microscopy? Although many might have had proteinase. Ku et al.3 present magnified analy- Properly controlled, however, refined expan- initial doubts, the new studies2–4 provide evi- sis of the proteome (MAP), which allows for sion microscopy techniques1–4,9 (Table 1) offer dence that expansion microscopy is a viable multiple rounds of immunostaining in an several advantages over ‘conventional’ super- and robust technology. expanded specimen. Repetitive rounds of resolution microscopy, including low cost Expansion microscopy relies on a flaw of immunostaining have previously been reported and the ease of performing multicolor experi- some tissue clearing protocols whereby the size for ultrathin sections7 and for cleared tissue8. ments. Another notable advantage is that thick

nature biotechnology volume 34 number 9 September 2016 929 news and views

specimens can be imaged with great detail in Research Council under the European Union’s 1. Chen, F., Tillberg, P.W. & Boyden, E.S. Science 347, all three spatial dimensions. Because refractive Seventh Framework Program (FP/2007-2013; ERC 543–548 (2015). 2. Tillberg, P.W. et al. Nat. Biotechnol. 34, 987–992 index heterogeneities of cells and tissues cause Grant Agreement n. 616791) and the German-Israeli Foundation. T.M. is also associated with the German (2016). optical aberrations, the spatial resolution of Center for Neurodegenerative Diseases (DZNE 3. Ku, T. et al. Nat. Biotechnol. 34, 973–981 (2016). 4. Chozinski, T.J. et al. Nat. Methods 13, 485–488 super-resolution microscopy quickly dete- Munich). P.E. was supported by the DFG Research (2016). riorates as the imaging plane is moved deeper Training Group 1373 and the Graduate School of 5. Galbraith, C.G. & Galbraith, J.A. J. Cell Sci. 124, into the specimen. This typically restricts the the Technische Universität München (TUM-GS); 1607–1611 (2011). C.F. is supported by the Graduate School of Systemic 6. Richardson, D.S. & Lichtman, J.W. Cell 162, 246–257 usefulness of super-resolution microscopy to Neuroscience (GSN, GSC 82). (2015). a few micrometers, which is often not enough 7. Micheva, K.D. & Smith, S.J. Neuron 55, 25–36 (2007). to capture the depth of one cell, let alone the 8. Murray, E. et al. Cell 163, 1500–1514 (2015). COMPETING FINANCIAL INTERESTS 9. Chen, F. et al. Nat. Methods 13, 679–684 (2016). architecture of complex tissues. In contrast, The authors declare no competing financial interests. 10. Dodt, H.U. et al. Nat. Methods 4, 331–336 (2007). specimens in expansion microscopy consist largely of water, so that optical aberrations are negligible and deep imaging is relatively straight forward. Precision medicine for Of course, expansion microscopy faces several challenges of its own. The samples can autoimmune disease become rather bulky, although objectives with extremely long working distances have already Lucienne Chatenoud been designed for cleared tissue, and re-slicing approaches (e.g., using an on-stage vibratome) An antigen-specific cell therapy for autoimmune disease avoids are conceivable. In addition, image brightness compromising immunity as a whole. diminishes as fluorophore density is diluted with expansion (i.e., by the third power of the expansion factor, as in all super-resolution approaches) The field of oncology is abuzz over recent degree to the development of all autoimmune and also from the necessary chemical treatments clinical trial results showing that patient diseases, injury of a given tissue usually results (up to a factor of two in the present reports2,4). T cells engineered in the laboratory are remark- from the predominant action of either one cell Thus, depending on the degree of desired resolu- ably effective against a few intractable cancers. type or the other. tion gain and the strength of chemical treatment Chimeric antigen receptor (CAR) T cells express Current treatments for autoimmune diseases needed for tissue homogenization, expansion a recombinant receptor that binds a specific are based on anti-inflammatory and immu- microscopy techniques might be restricted to tumor antigen, inducing the cell to kill target nosuppressive agents—including engineered relatively bright samples or might require addi- tumor cells. Writing in Science, Ellebrecht et al.1 biologics; human or humanized monoclo- tional amplification steps. Ameliorating this have now adapted the approach to autoimmune nal antibodies; and fusion proteins selective concern is the near-transparency of specimens, disease. They developed chimeric autoantibody for certain immune cell subsets or signaling which will facilitate imaging modalities geared receptor (CAAR) T cells as an antigen-specific pathways3—but their effect is transient and toward dim samples and fast imaging, such as therapy for the autoimmune disease pemphigus not antigen-specific. Chronic administration Nature America, Inc. All rights reserved. America, Inc. © 201 6 Nature light-sheet microscopy10. Moreover, the limits of vulgaris and showed, both in vitro and in mice, of these agents leads to the common side effects expansion microscopy in improving resolution the capacity of the cells to selectively eliminate of general immunosuppression, such as an remain to be determined. It is unclear whether B lymphocytes that produce autoantibodies to increased incidence of infections. npg specimens can be isotropically expanded tenfold desmoglein (Dsg) 3, the pathogenic mediators Major efforts have been devoted to selectively (or more) to reach the equivalent to 20-nm lateral of the disease. If these results can be translated targeting the autoantigen-specific response in resolution, now achievable with super-resolution to human autoimmune disease, this would various autoimmune diseases (multiple sclero- microscopy. Finally, as expansion microscopy represent a major leap toward achieving dura- sis, type 1 diabetes, and uveitis) by administering is a fixed-tissue method, it cannot be used on ble remissions or cures for certain severe and the autoantigen(s) through different routes (sub- live samples, an area in which super-resolution disabling conditions. cutaneous, oral, and parenteral). However, the microscopy is making steady progress. Autoimmune diseases can affect almost any efficacy seen in induced or spontaneous mouse All of this suggests that super-resolution micro- organ, and stem from a breakdown in immune models of autoimmunity has never been success- scopes and ‘ballooned’ samples will be allies, tolerance to host-derived or ‘self’ antigens. fully translated to the clinic. Among biological not competitors, in the search for finer cellular Their frequency has risen steadily over the past agents, CD3 antibodies have shown particular detail —yet another example of how physics and four decades in industrialized countries2, and promise for reversing established autoimmunity chemistry together shape biological inquiries. together they represent the third leading cause and durably restoring self-tolerance in animal of morbidity and mortality after cardiovascular models. Clinical development of this approach ACKNOWLEDGMENTS disease and cancer. Although both autoreac- is still in progress, but present data suggest The authors thank the Neurobiology Course at tive T and B lymphocytes contribute to some that combination therapies may be needed to the Marine Biological Laboratory for being their 3 summer home for many years. T.M.’s laboratory is achieve a sustained effect in humans . 1 supported by the Deutsche Forschungsgemeinschaft Ellebrecht et al. have approached the chal- (DFG) through the Munich Center for Systems Lucienne Chatenoud is at the Université Paris lenge of antigen-specific therapy by designing Neurology (SyNergy; EXC 1010), the Center for Descartes, Sorbonne Paris Cité, Paris, France, a modified form of a CAR that leads T cells to Integrated Protein Science Munich (CIPSM, EXC and INSERM U1151, CNRS UMR 8253, INEM kill autoreactive B cells. A conventional CAR 114), Collaborative Research Center 870, Priority Program 1710 and DFG Research Grant (MI Hôpital Necker-Enfants Malades, Paris, France. fusion protein consists of an extracellular anti- 674/7-1). Further support came from the European e-mail: [email protected] body moiety specific to an antigen of interest

930 volume 34 number 9 September 2016 nature biotechnology