Quantitative Visualization of Leukocyte Infiltrate in a Murine Model of Fulminant Myocarditis by Light Sheet Microscopy

Summary

Here, we describe a light-sheet microscopy approach to visualize the cardiac CD45⁺ leukocyte infiltrate in a murine model of aseptic fulminant myocarditis, which is induced by the intratracheal diphtheria toxin treatment of CD11c.DTR mice.

Abstract

Light-sheet fluorescence microscopy (LSFM), in combination with chemical clearing protocols, has become the gold standard for analyzing fluorescently labelled structures in large biological specimens, and is down to cellular resolution. Meanwhile, the constant refinement of underlying protocols and the enhanced availability of specialized commercial systems enable us to investigate the microstructure of whole mouse organs and even allow for the characterization of cellular behavior in various live-cell imaging approaches. Here, we describe a protocol for the spatial whole-mount visualization and quantification of the CD45⁺ leukocyte population in inflamed mouse hearts. The method employs a transgenic mouse strain (CD11c.DTR)that has recently been shown to serve as a robust, inducible model for the study of the development of fulminant fatal myocarditis, characterized by lethal cardiac arrhythmias. This protocol includes myocarditis induction, intravital antibody-mediated cell staining, organ preparation, and LSFM with subsequent computer-assisted image post-processing. Although presented as a highly-adapted method for our particular scientific question, the protocol represents the blueprint of an easily adjustable system that can also target completely different fluorescent structures in other organs and even in other species.

Introduction

During the evolution of light microscopy, many specialized forms appeared, all of them developed to minimize limitations in the visualization process for particular specimens. One such method is light-sheet fluorescence microscopy (LSFM). First developed in 1903 by Siedentopf and Zsigmondy¹ and finding its first fundamental biological applications in the early 1990s², LSFM has become the most powerful microscopic tool to date for the visualization of large specimens, such as intact mouse organs, with a fluorescence signal resolution down to the cellular level. Because of these benefits, in combination with its potential for live-cell imaging, Nature Methods named LSFM the "Method of the Year 2014³."

As the name suggests, the sample illumination in an LSFM is conducted by light sheets, which are orientated perpendicularly to the axis of the objective used for emission-light collection and subsequent image formation. The light sheet is usually generated either by focusing wide, collimated laser beams with a cylindrical lens or by the rapid sideways movement of narrow, focused laser beams in just one horizontal or vertical plane^4,5. In this way, only the focal plane of the imaging optics is illuminated, typically with a thickness of 1-4 µm. Consequently, for a fluorescent sample placed in the illumination plane, both the generation of scattered light and the effects of photobleaching from regions above or below the focal plane are eliminated or greatly suppressed^6,7. As all out-of-focus planes are not illuminated, photobleaching effects are omitted in these areas. In contrast to standard confocal or multi-photon microscopy, the paths of illuminated and emitted light are separate from each other, so the final image quality does not depend on the perfect focusing of the excitation light beam through the objectives. Depending on the underlying question, it is therefore possible to use objectives with an enormous field of view (FOV) so that the largest possible area of the illuminated plane can be imaged without any component parts moving in the xy-direction.

In modern LSFM systems, a fluorescent image of the generated optical section is captured on a highly sensitive charge-coupled device (CCD) or complementary metal-oxide semiconductor (CMOS) cameras, which are able to acquire the entire field of view (FOV) in microseconds. Therefore, by moving the sample through the light sheet and by acquiring images at defined z-steps, it is possible to obtain the full 3D information of a specimen in a reasonable amount of time^8,9, making this technique applicable for live-cell studies^10,11,12.

Nevertheless, although LSFM offers a rapid, sensitive, and fluorescence-friendly method, light transmission through the specimen is still a major issue, especially when large biological samples are the target for a 3D analysis. Light transmission is critically modified by physical aspects of absorption and by light scattering at interfaces of structures with different refractive indices¹³. Therefore, when imaging samples several millimeters in size, LSFM is mostly combined with clearing protocols to render the samples optically transparent. These techniques are based on the idea of removing water from the respective biological tissue and exchanging it with water- or (organic) solvent-based immersion media, which are chosen to narrowly match the refractive indices of the particular target tissue components. As a result, lateral light scattering is minimized, and all wavelengths of light can much more efficiently pass through the tissue¹³. In many cases, biological specimens treated in this way appear macroscopically crystal-clear, which enables LSFM to be conducted, even on entire mouse organs, using long working-distance, low magnification objectives.

Here, we present a preparation protocol for large-sample imaging in a light-sheet microscope (see the table of materials), which we have established to investigate the cellular cardiac infiltrate in a murine model of myocarditis¹⁵. CD11c.DTR mice express the primate diphtheria toxin receptor (DTR) under the control of the CD11c promotor¹⁶. Consequently, cells in these mice, which express CD11c along with the DTR, are rendered sensitive to the exotoxin of Corynebacterium diphtheria (diphtheria toxin, DTX); the systemic treatment of these animals with DTX results in a depletion of all CD11c⁺ cells. CD11c is an integrin and, as a cell-surface receptor for a variety of different soluble factors, is involved in activation and maturation processes mainly in cells of the monocytic lineage¹⁷. Consequently, the CD11c.DTR mouse model has been intensively used to study the role of dendritic cells and macrophage subsets in the context of many different immunological questions. Over time, it has been reported that CD11c.DTR mice treated systemically with DTX can develop adverse side effects and can display strongly elevated mortality rates^18,19. Recently, we were able to identify the underlying cause of death¹⁵, describing the development of fulminant myocarditis after intratracheal DTX application in these animals. The toxin challenge caused cellular destruction, including in central parts of the stimulus transmission system in the heart. This was accompanied by massive CD45⁺ leukocyte infiltrate, finally leading to lethal cardiac arrhythmias. In this case, not only was the appearance of the leukocyte population important, but also its spatial distribution inside the heart. This experimental question is a challenge for modern microscopic imaging, which we have solved by a light-sheet microscopy approach that is supported by intravital antibody staining and an organic solvent-based optical clearing protocol.

Protocol

All animal experiments were conducted in accordance with EU guidelines and were approved by the relevant local authorities in Essen (AZ 84-02.04.2014.A036 - Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, Essen, Germany). The animals were used and housed under specific pathogen-free (SPF) conditions.

1. Induction of Myocarditis by Diphtheria Toxin (DTX)

1. Prepare the DTX solution for the induction of myocarditis by diluting the DTX stock solution with phosphate-buffered saline (PBS) to a 1 µg/mL working solution.
2. Apply 100 µL of this solution intratracheally (i.t.) into 8 to 10 week-old CD11c.DTR mice (ca. 5 ng/g bodyweight (bw)), as shown by Hasenberg et al. for the i.t. application of a fungal spore suspension²⁰.
   NOTE: The intraperitoneal application of the toxin might result in a similar induction of fatal myocarditis. However, this approach would have to first be validated.
   1. For i.t. application, anesthetize the animals by an intraperitoneal (i.p.) injection of 100 mg/kg bw ketamine and 10 mg/kg bw xylazine.
   2. After reaching deep narcosis (toe pinch test), intubate the animals by using a 22 G indwelt venous catheter through the oral cavity. Check the correct positioning of the lens tube by ventilating the animals with a small-animal respirator at a rate of 250 breaths per min and an inhalation volume of 250 µL per breath, also shown by Hasenberg et. al.²⁰.
      NOTE: When positioned correctly, the breath rate of the animals will change according to the applied ventilation settings.
   3. Disconnect the ventilation system and instill the 100 µL of DTX solution or an equal amount of PBS as a control through the catheter into the lung and continue to ventilate the animals for 1 additional min.
3. Let the animals recover from anesthesia and monitor the general health status and body weight change for 4 days.
   NOTE: Depending on the genetic background, mouse strain, or initial weight, the severity and onset of myocarditis might vary and should be determined first.

2. Sample Preparation for Light-sheet Microscopy

1. 4 days after toxin application, anesthetize the mice by flushing an induction chamber with a vaporized 2% isoflurane/oxygen mixture. Inject 15 µg of a directly labeled anti-CD45 antibody (clone 30-F11, AlexaFluor647 (AF647)) in 50 µL of PBS intravenously (i.v.) using the retro-orbital application route, which ensures a very fast and highly accurate application of the correct amount of antibody.
   NOTE: The desired depth of narcosis has been reached when the mice are recumbent and do not react to toe pinch testing.
2. After 2 h of incubation, sacrifice the mice by cervical dislocation and perfuse the hearts in situ with 5 mL of PBS/EDTA (5 mM).
   1. Expose the heart and inject the perfusion solution into the right ventricle using a 21 G catheter. Perfuse with slow, constant pressure and make sure that the blood can be drained after cutting open the aorta.
      NOTE: This transcardial perfusion might result in minor preparation artifacts, which could disturb the final 3D visualization of the organ. Therefore, advanced animal experimenters could perform the perfusion through the inferior vena cava, excising the abdominal aorta.
3. For chemical fixation, perfuse the heart through the same recess with 5 mL of 4% paraformaldehyde (PFA). Keep the catheter in place and simply attach a new syringe filled with PFA. Perfuse with slow and constant pressure.
   NOTE: Paraformaldehydeis highly toxic; avoid contact with the skin, eyes, and other mucosa.
   1. Gently grasp the heart with serrated forceps; pull it slightly upwards; and cut all tissue connections, including the incoming and outgoing arteries and veins. Store the organ for an additional 4 h in 4% PFA for further fixation.
4. To dehydrate the organ, incubate the heart in an ascending ethanol series at 50%, 70%, and twice at 100%, each for 12 h, at 4 °C in the dark. Here, use black 5 mL polypropylene reaction tubes, allowing them to constantly rotate in a tube rotator using a slow rotation speed to prevent air-bubble formation.
5. Complete sample preparation by chemical clearing. To make the hearts transparent, incubate them in 98% dibenzyl ether (DBE) whilst constantly rotating overnight.
   NOTE: DBE is irritating to eyes, skin, and the respiratory system.

3. Light-sheet Microscopy

1. Conduct LSFM using the light-sheet microscope (see the table of materials). Place the sample on the standard sample holder of the microscope into the DBE-filled cuvette, which is suitable for specimens up to 30 mm x 30 mm x 15 mm.
2. Hold the sample firmly in place during the image acquisition using dehydrated and cleared agarose blocks.
   1. To prepare the agarose blocks, pour 2% molten agarose into a mold (15 mm x 15 mm x 5 mm). Let the agarose cool down until it solidifies.
   2. Dehydrate it in an ascending ethanol series (50%, 70%, and twice at 100%). Incubate the agarose blocks overnight in 98% DBE. Store the agarose in DBE until it is used.
      NOTE: As the agarose is comprised mainly of water, the incubation times for each dehydration step may be extended, depending on the size of the block.
      NOTE: DBE is irritating to the eyes, skin, and respiratory system.
   3. When needed, cut the prepared agarose block to the desired shape using a sharp scalpel. Typically, cut prism- or wedge-shaped forms to place at the edges of the sample adapter.
      NOTE: As the agarose blocks are elastic, the sample stays in place when pushed with little force.
3. Detect the fluorescence signals of interest (here, autofluorescence and anti-CD45 antibody staining) with the appropriate excitation and emission filter settings.
   1. For the detection of the connective tissue-derived autofluorescence signal, use a 525/50 nm bandpass filter at an excitation wavelength of 488 nm (OPSL: 50 mW); CD45 staining with an AF647-coupled antibody is detected at 668 nm (bandpass filter: 680/30 nm) and at an excitation wavelength of 647 nm (diode laser: 50 mW).
   2. Choose 4 µm as the distance between the two individual z-planes.

4. Image Post-processing

NOTE: The acquired digital image data were further processed with a scientific 3D/4D image processing and analysis software (see the table of materials).

1. Opening and pre-adjustment of the data files.
   1. After starting the analysis software, select the "Surpass" button in the upper-left corner. To open the image stacks, choose "File," "Open," and select a folder containing the data from the first acquired channel. Choose "Edit" and "Add Channels" to add further acquired channels.
      NOTE: Depending on the data size and the computer system, this process can take several minutes. The presented protocol demonstrates all steps for 2 acquisition channels (autofluorescence and AF647).
   2. Correct the x, y, and z voxel dimensions by clicking "Edit" and selecting "Image Properties." When the new window opens, adjust the parameters.
      NOTE: This step depends on the employed microscope system and the particular data format. Correct x-y-z values are critical for subsequent size and distant measurements. In most cases, these parameters are stored as meta data in the micrograph files. However, if the file format cannot be interpreted correctly by the analysis software, it is important to manually enter the pixel size. The pixel size in the x and y dimensions depend on the camera pixel size, the potentially applied binning, and the lens and objective magnification. The pixel size in z is equivalent to the self-defined z-step size (e.g., 4 µm).
2. Channel adjustments
   1. First transform the autofluorescence signal to grayscale by opening the "Display adjustment" ("Edit" and "Show Display Adjustment"). A new window will list all opened channels and the display settings for all of them; to change its default color, select the autofluorescence channel and choose the "white" display style. Click "ok" to apply.
   2. For black level adjustment, go back to "Display adjustment" and adjust the black level value to exclude unwanted background signals that do not represent sample structures.
      1. To do so, use the upper-left triangle on the channel bar and drag it from left to right.
         NOTE: Changes will be visualized immediately. Make sure to only suppress signals not deriving from the sample. Minimum values in the background should never be "0" to avoid pitch-black pixels, as the noise signals might be needed for additional calculations. Make sure that all black level changes are performed after image acquisition. Otherwise, valuable sample information may be lost. The black-level adjustment just serves representative purposes. To use the same display settings over different samples, precise values can be entered into the "min" numerical field below the channel list. This is very helpful for quantitative analyses.
   3. Perform intensity adjustment, either by using the upper-right triangle to adjust the respective channel intensity or by entering precise values into the "max" numerical field below the channel list.
   4. Perform contrast adjustment by dragging the lower triangle in the middle of the channel bar to increase or decrease the gamma value or by entering the precise values.
   5. Adjust the AF647 channel according to the autofluorescence channel. As a difference, set the visualization of the AF647 signal to a heat-map color look-up table. Do this by choosing the channel mode "fire" in an approach similar to that in step 4.2.1.
      NOTE: As the overall signal intensities are significantly lower compared to the autofluorescence channel, the black levels are usually adjusted to much lower values. As for the autofluorescence channel, the brightness of this channel is usually increased.
   6. Select the "blend mode" to visualize tissue structures in detail. Analyze all samples from one series with the same parameter values.
3. Virtual clipping
   1. To virtually cut open the organ before the 3D scene export, apply a clipping plane on the rendered 3D object.
      1. Click the "Clipping Plane" icon (scissors) in the object list. Inside the image view, a yellow frame and a manipulator (white spindle) will appear. Rotate the spindle at the thinner end with the mouse, thereby changing the orientation of the clipping plane, and move the thicker middle part of the spindle to select the desired depth of clipping.
      2. Hide the frame and manipulator by unchecking the respective checkboxes below the object list and create the desired snapshots.

Results

The presented LSFM approach analyzes the leukocyte distribution and amount in murine hearts upon induction of severe myocarditis. Figure 1A explains the pre-treatment protocol for the transgenic CD11c.DTR mice for myocarditis induction. This step represents the necessary trigger for the recruitment of leukocytes to the myocardium. After successful DTX application, the animals develop severe disease symptoms, such as general weakness, anorexia, and weight loss within the r...

Discussion

In modern life science, the microscopic visualization of biological processes plays an increasingly important role. In this context, many developments have been achieved during the last two centuries that help to answer questions not addressable up to this point, . First, there has been a clear tendency to fundamentally increase the resolution ability of light microscopes. With structured illumination microscopy (SIM)^21,22, stimulated emission-depletion (STED) mi...

Materials

Name	Company	Catalog Number	Comments
diphtheria toxin	Sigma Aldrich	D0564
phosphate buffered saline	Biochrom	L182-10
ketamine [50 mg/mL]	Inresa	PZN 4089014
xylazine [2 %]	Ceva
syringe	Braun	REF 9161502	Braun Omincan F
indwelt venous catheter	Becton Dickinson	REF 381923	BD Insyte Autoguard
small animal respirator	Harvard Apparatus	73-0044	MiniVent
anit-CD45 antibody (AF647)	BioLegend	103124
ethylenediaminetetraacetic acid (EDTA) disodium salt dihydrate	Carl Roth	8043.2
catheter (21 G)	BD Biosciences	REF 387455	BD valu-set
paraformaldehyde	Sigma Aldrich	P6148
PE tube (5 mL)	Carl Roth	EKY9.1	Rotilabo
ethanol	Carl Roth	9065.4
dibenzyl ether	Sigma-Aldrich	108014
tube rotator	Miltenyi Biotec	130-090-753	MACSmix
agarose (low gelling)	Sigma Aldrich	A4018
mold (15 x 15 x 5 mm)	Tissue Tek	4566	Cryomold
light sheet microscope system	LaVision Biotec		Ultramicroscope
microscope body	Olympus		MVX10
objective	Olympus		0.5 NA MVPLAPO 2XC, WD 5.5 mm
sCMOS camera 5.5MP	Andor Technology
488 nm OPSL (50mW) laser	Coherent
647 nm  diode laser (50mW)	Coherent
3D image processing & analysis software	Bitplane		IMARIS Ver. 8.3
transgenic mouse strain	Steffen Jung et al.		CD11c.DTR
wt mouse strain	Envigo		Balb/c Ola Hsd J
Laser Module	LaVision Biotec
MVPLAPO 2XC Objective Lens