Label-free, High-resolution Fluorescence Imaging of the Intestinal Wall by Image Scanning and Confocal Microscopy

Long considered the microscopist’s foe, autofluorescence (AF) has unveiled its potential as a source of image contrast in biological tissue. Several studies reported label-free imaging of the intestinal wall, but a detailed characterization of the different cell populations encountered in the four histological layers of this organ has been missing.1, 2, 3 We recently proposed an approach that enabled a fast and three-dimensional AF imaging of the enteric nervous system across the entire intestinal wall, using two-photon excitation at 800-nm and broad fluorescence collection. For the first time, this intricate meshwork comprised of nervous ganglia was resolved in its full three-dimensional complexity at a subcellular resolution and without any staining.⁴ However, the subcellular resolution was diffraction-limited and the signal-to-noise ratio (SNR) remained low compared to images obtained with classical histological stains. In the present work, we therefore extend our original strategy and we demonstrate linear AF imaging upon 405-nm excitation of mouse intestinal slices reaching 170–180-nm lateral resolution using image-scanning microscopy (ISM)⁵ on a commercially available instrument.

Mouse intestines were fixed overnight at 4 °C in a 4% formaldehyde solution and processed using standard protocols.⁴ A 4-μm thin slice was then deparaffined using Histo-Clear, mounted with Dabco (25g/L) in a 90% glycerol/10% Tris 0.2M solution and imaged on a Zeiss LSM880 image-scanning confocal microscope equipped with a 40X/1.2 numerical aperture (NA) water objective. Three images were acquired for each region-of-interest: (i) a conventional laser-scanning confocal image upon 405-nm excitation; 0.015 mW (0,5%) laser power; detection in the 410- to 695-nm band for broad AF collection; (ii) a spectral image upon 405-nm excitation; 0.06 mW (2%) laser power; 410-695-nm AF collection split into 32 spectral channels; and finally, (iii) an AiryScan image again upon 405-nm excitation; 0.3 mW (10%) laser power; 420–445-nm and 465–505-nm AF collection. Images were postprocessed using Zeiss ZEN 3.7 and ImageJ.

Using Abbe's formula and the Rayleigh criterion, the smallest resolvable distance (diffraction limit) in the focal plane of a confocal microscope with the pinhole set to 1 Airy Unit can be estimated as d = (0,4^.λ)/NA where λ and NA are the wavelength and NA of the objective, respectively. In practice, resolutions of the order of 220–250-nm are attained in biological tissue at best, and it is important to note that spatial resolution and image contrast go hand in hand. For AF imaging of large fields of view (Figure A, left), this resolution is sufficient to image across the whole intestinal wall and to distinguish the 4 tissue layers and even some subcellular structures. However, when zooming in on the picture, tiny subcellular structures appear blurry, as a result of both the limited resolution and a poor SNR of the relatively faint and indiscriminate AF (Figure A, left, inset).

Figure.

(A) Confocal (left) and AiryScan (right) autofluorescence (AF) images of the mouse intestinal wall (ileum). Inset: magnification of granules in Paneth cells. Scale-bar: 20μm. (B) Spectral image in the same region-of-interest as shown in A. Inset: AF spectrum of the region shown as a yellow-dotted square. (C) Confocal (top) and AiryScan (bottom) AF images of a Paneth cell at the bottom of an intestinal crypt; (D) Confocal (top) and AiryScan (bottom) AF images of a granule from a Paneth cell and corresponding fluorescence intensity profiles highlighting the lower SNR in confocal images. (E and F) AiryScan AF images of a part of an intestinal villus (E) and a myenteric plexus ganglion (F). Scale-bars: 5 μm. Au, Auerbach (myenteric) plexus ganglia; bb, brush border; Ent, enterocyte; Gc, Goblet cell; l, lumen; m, mucosa; m-c, muscularis circular; m-l, muscularis longitudinal; n, nucleus; s, serosa; sg, secretory granules; sm, submucosa.

ISM is an easy and straightforward approach for improving contrast and resolution compared to standard confocal imaging.⁵ ISM consists of replacing the confocal pinhole and bucket detector by a small imaging array detector. Upon scanning, the image of the diffraction pattern is scanned over the detector. In the commercial "AiryScan" module (ZEISS) this detector is a 32-element photo multiplier array organized in a hexagonal honeycomb array. Each element acts as a virtual pinhole of 0.2 Airy units and the 32 microlenses cover an area that is 1,25 Airy units wide,⁶ improving the otherwise low light-efficiency of confocal detection, where only the central peak of the Airy pattern is captured.

AiryScan AF imaging increased the SNR allowing the observation of tiniest intracellular details (Figure A, right, inset). The improved ISM image is the result of a two-step process: (i) a deconvolution step using noniterative Wiener filtering over all phases reduces noise in each individual-phase contribution and (ii) a Shepard sum reassigns photons to the right pixels for reconstructing the final super-resolution image. The overall result is a reduction of the dynamic range compared to a standard confocal image, but an improved SNR (Figure D). Spectral imaging reveals a dominantly blue-green and spectrally fairly uniform AF over the entire intestinal walls (Figure B). AF in this spectral range mainly arises from cytokeratins, flavins, and lipopigment.^7,8 As already observed earlier in our study using a custom microscope and nonlinear excitation,⁴ contrast is thus generated mainly through variations in AF intensities among the biological structures rather than by specific spectral signatures. Applied to intestinal wall structures, AiryScan AF imaging allowed us the identification and detailed investigation of Paneth cells⁹ at the bottom of the crypts, with their characteristic secretory granules resolved (Figure C and D). At the level of the mucosa, both enterocytes and their brush border on the luminal side as well as goblet cells are clearly identified (Figure E). Enteric nervous system neurons from the myenteric (Auerbach) plexus can be observed in their environment, rich in muscular cells (Figure F). Comparison in between conventional confocal (Figure C, top) and Airyscan images (Figure C, bottom) confirms our earlier observation of a higher SNR. ISM thus provides sharper and more detailed images for the in-depth morphological characterization of the intestinal wall. More images can be seen on Figure A1.

In comparison to conventional confocal laser scanning microscopy, Airyscan requires longer acquisition times and higher laser power (14 seconds vs 1 minute 52 seconds and 0.015 mW vs 0.3 mW, respectively, for the images shown). Such parameters can induce photobleaching when repeatedly imaging the same region-of-interest (Figure A2). However, in our study, intestinal AF proved quite robust and the net increase of the SNR enables the obtention of sharp and detailed, label-free images that largely outweighs the higher laser power.

In summary, AF imaging generates images with similar information content as the ones obtained with traditional histological stains such as hematoxylin/eosin.¹⁰ However, these stains can only be observed with classic bright-field microscopy, and thus it is impossible to obtain either optical sectioning or images below the diffraction limit, both of which are possible with our approach. We found that AF signals arising from intestinal structures emit predominantly in the blue-turquoise wavelength range and thus at lower wavelengths than typical fluorescence labels. AF contrast can thus be used as a “counterstain" to provide contextual information in immunofluorescence experiments using probes in the orange/red special range. Finally, being label-free, AF imaging is fast and simple: once the slice or biopsy is dissected and fixed, the sample is mounted on the microscope and high-resolution images are obtained in less than 2 minutes.

AF ISM fills the gap in between conventional optical and electron microscopy. Our approach does not require the long, costly and tedious protocols of the later and it offers images of the various cellular and subcellular structures of the intestinal wall without any staining at a stunning detail. With its speed and ease, we anticipate applications in the gastrointestinal field, especially for diagnosis such as extemporaneous examination.