Special issue on high-resolution optical imaging

The pace of development in the field of advanced microscopy is truly breath-taking, and is leading to major breakthroughs in our understanding of molecular machines and cell function. This special issue of Journal of Optics draws attention to a number of interesting approaches, ranging from fluorescence and imaging of unlabelled cells, to computational methods, all of which are describing the ever increasing detail of the dynamic behaviour of molecules in the living cell. This is a field which traditionally, and currently, demonstrates a marvellous interplay between the disciplines of physics, chemistry and biology, where apparent boundaries to resolution dissolve and living cells are viewed in ever more clarity. It is fertile ground for those interested in optics and non-conventional imaging to contribute high-impact outputs in the fields of cell biology and biomedicine. The series of articles presented here has been selected to demonstrate this interdisciplinarity and to encourage all those with a background in the physical sciences to 'dip their toes' into the exciting and dynamic discoveries surrounding cell function.

Although single molecule super-resolution microscopy is commercially available, specimen preparation and interpretation of single molecule data remain a major challenge for scientists wanting to adopt the techniques. The paper by Allen and Davidson [1] provides a much needed detailed introduction to the practical aspects of stochastic optical reconstruction microscopy, including sample preparation, image acquisition and image analysis, as well as a brief description of the different variants of single molecule localization microscopy. Since super-resolution microscopy is no longer restricted to three-dimensional imaging of fixed samples, the review by Fiolka [2] is a timely introduction to techniques that have been successfully applied to four-dimensional live cell super-resolution microscopy. The combination of multiple high-resolution techniques, such as the combination of light sheet and structured illumination microscopy (SIM), which efficiently utilize photon budget and avoid illuminating regions of the specimen not currently being imaged, hold the greatest promise for future biological applications. Therefore, the combined setup for SIM and single molecule localization microscopy (SMLM) described by Rossberger et al [3] will be very helpful and stimulating to advanced microscopists in further modifying their setups. The SIM image helps in identifying artefacts in SMLM reconstruction, e.g. when two active fluorophores are close together and get rejected as 'out-of-focus'. This combined setup is another way to facilitate imaging live samples. The article by Thomas et al  [4] presents another advance for biological super-resolution imaging with a new approach to reconstruct optically sectioned images using structured illumination. The method produces images with higher spatial resolution and greater signal to noise compared to existing approaches. This algorithm demonstrates great promise for reconstructing biological images where the signal intensities are inherently lower.

Shevchuk et al [5] present a non-optic near field approach to imaging with a review of scanning ion-conductance microscopy. This is a powerful alternative approach for examining the surface dynamics of living cells including exo and endocytosis, unlabelled, and at the level of the single event. Here they present the first data on combining this approach with fluorescence confocal microscopy—adding that extra dimension. Different approaches to label-free live cell imaging are presented in the papers by Patel et al  [6], Mehta and Oldenbourg [7], as well as Rogers and Zheludev [8]. All three papers bring home the excitement of looking at live cell dynamics without reporters—Patel et al  [6] review both the potential of coherent anti-Stokes Raman scattering and biological applications, where specific biomolecules are detected on the basis of their biophysical properties. Polarized light microscopy as presented by Mehta and Oldenbourg [7], describe a novel implementation of this technology to detect dichroism, and demonstrate beautifully its use in imaging unlabelled microtubules, mitochondria and lipid droplets. Sub-wavelength light focusing provides another avenue to super-resolution, and this is presented by Rogers and Zheludev [8]. Speculating on further improvements, these authors expect a resolution of 0.15λ. To date, the method has not been applied to low contrast, squishy and motile biotargets, but is included here for the clear potential to drive label-free imaging in new directions. A similar logic lies behind the inclusion of Parsons et al  [9] where ultraviolet coherent diffractive imaging is further developed. These authors have demonstrated a shrink-wrap technique which reduces the integration time by a factor of 5, bringing closer the time when we have lab based imaging systems based on extreme ultraviolet and soft x-ray sources using sophisticated phase retrieval algorithms.

Real biological specimens have spatially varying refractive indices that inevitably lead to aberrations and image distortions. Global refractive index matching of the embedding medium has been an historic solution, but unfortunately is not practical for live cell imaging. Adaptive optics appears an attractive solution and Simmonds and Booth [10] demonstrate the theoretical benefits of applying several adaptive optical elements, placed in different conjugate planes, to create a kind of 'inverse specimen' that unwarps phase distortions of the sample—but these have yet to be tested on real specimens. A difficulty in single molecule localization microscopy has been the determination of whether or not two molecules are colocalized. Kim et al  [11] present a method for correcting bleed-through during multi-colour, single molecule localization microscopy. Such methods are welcome standards when trying to quantifiably interpret how close two molecules actually are. Rees et al  [12] provide an invaluable overview of key image processing steps in localization microscopy. This paper is an excellent starting point for anyone implementing localization algorithms and the Matlab software provided will be invaluable; a strong paper on which to conclude our overview of the excellent articles brought together in this issue.

One aspect brought home in several of these articles is the volume of data now being collected by high resolution live cell imaging. Data processing and image reconstruction will continue to be pressure points in the further development of instrumentation and analyses. We would hope that the series of papers presented here will motivate software engineers, optical physicists and biologists to contribute to the further development of this exciting field.

References

[1] Allen J R et al 2013 J. Opt. 15 094001

[2] Fiolka R et al 2013 J. Opt. 15 094002

[3] Rossberger S et al 2013 J. Opt. 15 094003

[4] Thomas B et al 2013 J. Opt. 15 094004

[5] Shevchuk A et al 2013 J. Opt. 15 094005

[6] Patel I et al 2013 J. Opt. 15 094006

[7] Mehta S B et al 2013 J. Opt. 15 094007

[8] Rogers E T F et al 2013 J. Opt. 15 094008

[9] Parsons A D et al 2013 J. Opt. 15 094009

[10] Simmonds R et al 2013 J. Opt. 15 094010

[11] Kim D et al 2013 J. Opt. 15 094011

[12] Rees E J et al 2013 J. Opt. 15 094012

In recent years there has been a rash of developments in light microscopies circumventing traditional resolution limits associated with the diffraction of light occurring between the sample and the detector. Collectively, these techniques are referred to as 'superresolution' microscopies. One major family of superresolution techniques, variably referred to as PALM, FPALM and STORM, uses temporal control of the excited state of fluorophores to sequentially identify single non-overlapping emitters in time and space. Conventional images of single point emitters are fitted to sub-diffraction-limited areas, and a composite image is reconstructed from position data collected over many thousands of individual imaging frames. This paper provides a brief overview of superresolution microscopy, followed by a detailed discussion of STORM, including practical guidelines for sample preparation designed to help to make the technique more accessible to the non-specialist.

In fluorescence microscopy it has become possible to fundamentally overcome the diffraction limited resolution in all three spatial dimensions. However, to have the most impact in biological sciences, new optical microscopy techniques need to be compatible with live cell imaging: image acquisition has to be fast enough to capture cellular dynamics at the new resolution limit while light exposure needs to be minimized to prevent photo-toxic effects. With increasing spatial resolution, these requirements become more difficult to meet, even more so when volumetric imaging is performed. In this review, techniques that have been successfully applied to three-dimensional, super-resolution live microscopy are presented and their relative strengths and weaknesses are discussed.

Understanding the positional and structural aspects of biological nanostructures simultaneously is as much a challenge as a desideratum. In recent years, highly accurate (20 nm) positional information of optically isolated targets down to the nanometer range has been obtained using single molecule localization microscopy (SMLM), while highly resolved (100 nm) spatial information has been achieved using structured illumination microscopy (SIM).

In this paper, we present a high-resolution fluorescence microscope setup which combines the advantages of SMLM with SIM in order to provide high-precision localization and structural information in a single setup. Furthermore, the combination of the wide-field SIM image with the SMLM data allows us to identify artifacts produced during the visualization process of SMLM data, and potentially also during the reconstruction process of SIM images.

We describe the SMLM–SIM combo and software, and apply the instrument in a first proof-of-principle to the same region of H3K293 cells to achieve SIM images with high structural resolution (in the 100 nm range) in overlay with the highly accurate position information of localized single fluorophores. Thus, with its robust control software, efficient switching between the SMLM and SIM mode, fully automated and user-friendly acquisition and evaluation software, the SMLM–SIM combo is superior over existing solutions.

A new approach to reconstructing an optically sectioned image using structured illumination is presented. Compared to the algorithm proposed by Neil et al (1997 Opt. Lett. 22 1905), this method uses the same number of images to construct a final image with a flat linear transfer function at the cost of slightly more complexity in the reconstruction algorithm. Compared to other optical sectioning algorithms using structured illumination, this approach produces images with higher contrast and better image fidelity at low signal intensities.

Optical visualization of nanoscale morphological changes taking place in living biological cells during such important processes as endo- and exocytosis is challenging due to the low refractive index of lipid membranes. In this paper we summarize and discuss advances in the powerful combination of two complementary live imaging techniques, ion conductance and fluorescence confocal microscopy, that allows cell membrane topography to be related with molecular-specific fluorescence at high spatial and temporal resolution. We demonstrate the feasibility of the use of ion conductance microscopy to image apical plasma membrane of mouse embryo trophoblast outgrowth cells at a resolution sufficient to depict single endocytic pits. This opens the possibility to study individual endocytic events in embryo trophoblast outgrowth cells where endocytosis plays a crucial role during early stages of embryo development.

Coherent anti-Stokes Raman scattering (CARS) has established itself as an imaging technique capable of providing video-rate imaging of biological specimens through vibrational coherence of endogenous molecules. Current techniques predominantly involve the application of costly, invasive and potentially non-specific dyes or labels for imaging biomolecules. CARS microscopy can however provide a high-resolution and non-invasive alternative for imaging biomolecules of interest without the need for exogenous labels. Here we provide an overview of CARS including the technique and common instrumentation as well as its applications in biomedical imaging. We discuss the major biomedical areas where CARS has been applied such as in evaluating liver disease, progression of atherosclerosis, tumour classification and tracking drug delivery, whilst also assessing the future challenges for clinical translation.

Polarized light microscopy provides unique opportunities for analyzing the molecular order in man-made and natural materials, including biological structures inside living cells, tissues, and whole organisms. 20 years ago, the LC-PolScope was introduced as a modern version of the traditional polarizing microscope enhanced by liquid crystal devices for the control of polarization, and by electronic imaging and digital image processing for fast and comprehensive image acquisition and analysis. The LC-PolScope is commonly used for birefringence imaging, analyzing the spatial and temporal variations of the differential phase delay in ordered and transparent materials. Here we describe an alternative use of the LC-PolScope for imaging the polarization dependent transmittance of dichroic materials. We explain the minor changes needed to convert the instrument between the two imaging modes, discuss the relationship between the quantities measured with each instrument, and touch on the physical connection between refractive index, birefringence, transmittance, diattenuation, and dichroism.

Optical super-oscillations, first predicted in 1952 and observed in 2007, offer a promising route to optical super-resolution imaging and show potential for manufacturing with light and data-storage applications such as direct optical recording and heat assisted magnetic recording. We review the history and basic physics behind the phenomenon of super-oscillation and its application in optics. We overview recent results in creating optical super-oscillations using binary masks, spatial light modulators and planar metamaterial masks. We also investigate the limits and competitiveness of super-oscillatory imaging.

Conventional coherent diffractive imaging (CDI) techniques rely on inversion of the two-dimensional phase problem in the fully coherent limit. Current work using synchrotrons has shown that by introducing a flexible parameter in a technique known as polyCDI, some reduced temporal coherence with relative bandwidth ∼3% can be tolerated for simple non-dispersive objects. We demonstrate that using a high harmonic source with modulated spectral characteristics, although the excellent temporal coherence properties are lost in detection, it is possible to increase the tolerable relative bandwidth to ∼20% by using the shrinkwrap technique and treating the data as if they were fully coherent but noisy. This reduces the integration time by a factor of ∼5. This result is critical for the future use of lab-based sources of extreme ultraviolet and soft x-rays for CDI of non-dispersive objects, and we anticipate that it will improve results at synchrotron sources also.

Adaptive optics has been implemented in a range of high-resolution microscopes in order to overcome the problems of specimen-induced aberrations. Most implementations have used a single aberration correction across the imaged field. It is known, however, that aberrations often vary across the field of view, so a single correction setting cannot compensate all aberrations. Multi-conjugate adaptive optics (MCAO) has been suggested as a possible method for correction of these spatially variant aberrations. MCAO is modelled to simulate the correction of aberrations, both for simple model specimens and using real aberration data from a biological specimen.

Multi-colour localization microscopy has enabled sub-diffraction studies of colocalization between multiple biological species and quantification of their correlation at length scales previously inaccessible with conventional fluorescence microscopy. However, bleed-through, or misidentification of probe species, creates false colocalization and artificially increases certain types of correlation between two imaged species, affecting the reliability of information provided by colocalization and quantified correlation. Despite the potential risk of these artefacts of bleed-through, neither the effect of bleed-through on correlation nor methods for its correction in correlation analyses have been systematically studied at typical rates of bleed-through reported to affect multi-colour imaging. Here, we present a reliable method of bleed-through correction applicable to image rendering and correlation analysis of multi-colour localization microscopy. Application of our bleed-through correction shows that our method accurately corrects the artificial increase in both types of correlation studied (Pearson coefficient and pair correlation), at all rates of bleed-through tested, in all types of correlation examined. In particular, anti-correlation could not be quantified without our bleed-through correction, even at rates of bleed-through as low as 2%. While it is demonstrated with dichroic-based multi-colour FPALM here, our presented method of bleed-through correction can be applied to all types of localization microscopy (PALM, STORM, dSTORM, GSDIM, etc), including both simultaneous and sequential multi-colour modalities, provided the rate of bleed-through can be reliably determined.

Localization microscopy software generally contains three elements: a localization algorithm to determine fluorophore positions on a specimen, a quality control method to exclude imprecise localizations, and a visualization technique to reconstruct an image of the specimen. Such algorithms may be designed for either sparse or partially overlapping (dense) fluorescence image data, and making a suitable choice of software depends on whether an experiment calls for simplicity and resolution (favouring sparse methods), or for rapid data acquisition and time resolution (requiring dense methods). We discuss the factors involved in this choice. We provide a full set of MATLAB routines as a guide to localization image processing, and demonstrate the usefulness of image simulations as a guide to the potential artefacts that can arise when processing over-dense experimental fluorescence images with a sparse localization algorithm.