Animals and hematopoietic cell transplantation.
BALB/c mice (H-2d, Thy1.2) and C57BL/6 mice (H-2b, Thy1.2) were purchased from Charles River at the age of 8–16 weeks. Transgenic mice C57BL/6.L2G85 CD90.1+
expressing luciferase (luc+
) were generated by backcrossing the luc+
FVB/N-L2G85 founder line (14
) with the C57BL/6 mice for more than 12 generations. Additional transgenic luc+
mice C57BL/6.L2G85 CD45.1+
mice were bred in our colony. BALB/c mice were conditioned with a total body irradiation dose of 8 Gy and i.v. injected with 5 × 106
C57BL/6 WT sex- and age-matched bone marrow cells within 2 hours after irradiation. C57BL/6.L2G85 CD90.1+
magnetic bead– separated (Invitrogen) T cells (CD4/CD8 enrichment, approximately 80% purity) were coinjected to induce GVHD.
BLI was performed as previously described (13
) using an IVIS Spectrum CCD imaging system (Caliper Xenogen). Mice were anesthetized with an intraperitoneally injected mixture of ketamine (50 μg/g body weight, BW) and xylazine (5 μg/g BW) in 0.1 M PBS in a total volume of 10 μl/g BW. d
-Luciferin was injected at a concentration of 150 μg/g BW. Imaging data were analyzed with Living Image 4.0 (Caliper Xenogen).
The following murine antibodies were purchased from BioLegend and Invitrogen: CD4 (GK1.5), CD4 (RM4-5), CD8α (clone 53-6.7), CD11c (N418), CD31 (WM59), CD35 (7E9), CD44 (IM7), CD62L (MEL-14) CD90.1 (OX-7), CD45.1 (A20), CD105 (MJ7/18), CD106 (429/MVCAM.A), MAdCAM-1 (MECA-367), and podoplanin (clone 8.1.1). Human antibodies were purchased from BioLegend: CD4 (OKT4), CD8α (HIT8a), and CD31 (WM59). FACS and standard immunofluorescence microscopy were performed using an Imager Z1M (Zeiss) and a BD FACSCanto II flow cytometer (BD Biosciences).
Magnetic bead–enriched T cells were further sorted into CD4+CD44hiCD62Llo TEM, CD4+CD44hiCD62Lhi TCM, and CD4+CD44loCD62Lhi TN using a FACSAria II cell sorter (BD Biosciences).
Preparation of mouse specimens.
Organs were prepared for LSFM by modified protocols of previously described procedures (7
). Briefly, mice were anesthetized by intraperitoneal injection of ketamine-xylazine and transcardially perfused with 20 ml ice-cold PBS, followed by 40 ml of 4% paraformaldehyde (pH 7.4). Following the perfusion, organs were removed. The samples were placed and stored in paraformaldehyde for at least 2 hours at 4°C until use in further procedures. Hemoglobin-rich organs were bleached for 30 minutes in 15% hydrogen peroxide/methanol. For homogenous ex situ immunofluorescence staining of large specimens, the tissue samples were blocked for 18–24 hours with 2% FCS/PBS in 0.1% Triton-X and afterward incubated with the respective antibodies for 24 hours at 4°C on a shaker, washed in PBS, incubated with streptavidin for 24 hours, and washed again in PBS, followed by dehydration in a graded ethanol series (30%, 50%, 70%, 80%, 90%, 96%, and in 100% for 2 hours each) at room temperature. After the samples were rinsed for 2 hours in 100% n
-hexane, the n
-hexane was replaced stepwise by a clearing solution consisting of 1 part benzyl alcohol in 2 parts benzyl benzoate (Sigma-Aldrich). Air exposure was strictly avoided at this step. Tissue specimens became optically transparent and suitable for the LSFM imaging after incubation in the clearing solution for at least 2 hours at room temperature.
In vivo antibody staining.
Fluorochrome-labeled antibodies were injected i.v. 2.5 hours (300 μg CD4, CD45.1, CD90.1, or CD11c) or 0.5 hours (50 μg MAdCAM-1) before mice were euthanized. To confirm distribution and specificity of i.v. antibody staining, additional post-section staining was performed with the same antibody. To demonstrate antibody distribution and specificity of inflamed and non-inflamed tissues, CD4–Alexa Fluor 647 antibody was injected into untreated mice or into mice on day +6 after allo-HCT. Two and a half hours later, mice were euthanized, and spleens and PPs were harvested and digested according to a previously described protocol (32
) with slight modifications. Briefly, organs were digested (2 mg/ml collagenase D and 0.1 mg/ml DNase I) at 37°C for 30 minutes in the presence of saturating amounts of CD4-FITC antibody (30 μg/ml) to prevent ex situ staining of remaining unbound CD4–Alexa Fluor 647. Thus, CD4-FITC single-positive cells indicate ex situ stained cells. If not specifically indicated as antibody i.v. staining, fluorescence staining was performed according to the ex situ staining protocol.
Preparation of human tissue.
The preparation of human tissue required important modifications. Fresh biopsy tissue was fixed for 2–4 hours in PFA, washed in PBS, bleached in 15% hydrogen peroxide/methanol, and again washed in PBS. Subsequently, the sample was pretreated with 0.1% Triton-X/PBS for 24 hours before addition of 2% FCS for another 8 hours. To allow for homogenous staining of deeper tissue layers, antibody incubation was performed for 4–6 days in 0.1% Triton-X/PBS. After washing again in PBS, the sample was dehydrated in a graded ethanol series for 4 hours each. The clearing procedure was performed exactly as for the preparation of mouse tissue.
We used an LSFM setup similar to a previously described setup (8
); however, the sample was illuminated from one side by various laser sources: 405-nm diode laser (56ICS025, CVI Melles Griot), 473-nm DPSS laser (MBL-473 100 mW, CNI), 532-nm DPSS laser (MGL-W532 500 mW, CNI), and 639-nm diode laser (Cube 640-40, Coherent). An objective inverter (LSM Tech) on a commercial inverted microscope (Axiovert 200, Zeiss) allowed a flexible horizontal positioning of the objective at a greater distance and range than typically possible with an inverted microscope. The laser excitation beams (3–4 mm diameter) were expanded 3-fold with a telescope (Thorlabs). With focus beam waist of 4–9 μm, an approximately 1-cm-tall light sheet was created by focusing the beam in a cylindrical lens (f
= 5 cm and f
= 3.6 cm, Newport) depending on the detecting objective lens (Apochromat 5× with 0.16 numerical aperture [NA] and LD Achroplan 20× with 0.4 NA objectives [Zeiss]). The beam light sheet was adjusted exactly along the focus plane of a microscope. Fine focusing of the light sheet on the specimen was performed by moving the cylindrical lens mounted on a translation stage (Standa).
After sample fixation and clearing, the specimen was affixed with Pattex acrylamide glue (Henkel) to a home-built glass rod and placed in a home-built coverglass chamber filled with clearing solution. The sample was positioned with a stepper motor–controlled stage (Standa). The stage with a feedback memory system enabled translational and rotational step motion with accuracy of 0.1 μm and 0.1 degrees, respectively.
The fluorescence emission light was filtered using a motorized filter wheel (Standa) according to the excitation wavelength: DAPI 405 nm HQ440/40, Alexa Fluor 488 (Invitrogen) 488 nm HQ525/50, Alexa Fluor 532 (Invitrogen) HQ605/75, Alexa Fluor 647 (Invitrogen) 640 nm HQ695/55 (all Chroma). Fluorescence was detected by a back-illuminated electron multiplying charge-coupled device (EMCCD) camera (512 × 512 pixels, 16 μm/pixel Cascade II, Photometrics). Exposure times for image acquisition were 200–500 ms per frame. To acquire the complete image stack, each excitation wavelength and emission filter combination was used sequentially plane by plane. We used the software package LabView (National Instruments) to control all hardware, including the synchronization with data acquisition software MetaMorph 7.1 (Molecular Devices). Finally, the individual color stacks were overlaid via image processing using the Volocity software package (PerkinElmer).
Image acquisition and processing.
We acquired multicolor stacks, imaging each plane sequentially by each wavelength and emission filter combination. Stacks were taken in increments of either 1, 2, or 5 μm. Exposure times for image acquisition were 200 or 500 ms per frame. The excitation wavelengths, emission filter colors, data acquisition, and storage were synchronized by LabView software (National Instruments) and administered using the software package MetaMorph 7.1 (Molecular Devices). Resulting multicolor stacks were processed, if necessary (e.g., cropped or reduced), with ImageJ (NIH). Subsequently, we used the 3D image processing software Volocity (PerkinElmer) to prepare the individual images and videos.
Samples were embedded into a self-constructed chamber for MPM. The MPM was equipped with an optical parametric oscillator (OPO, APE) for two-photon irradiation at 1,100 nm. The excitation beam was focused with a ×20 NA 0.95 water objective (Olympus). Emission was detected with HQ535/50-nm, HQ605/70-nm, and HQ710/75-nm filters. Sequential 3D stacks were obtained for up to 1-mm penetration depth at a step size of 5 μm. Single-color image stacks (overlap, 16%) were stitched together, image size was reduced from 5,566 × 3768 pixels to 756 × 512 pixels, and photomultiplier tube–induced (PMT-induced) noise outliers were removed with ImageJ (processing package Fiji, NIH) before 3D image processing with Volocity.
Volume calculations and automated cell counting.
Multicolor stacks were processed on Volocity software (Improvision) for 3D rendering and subsequently volume quantification and cell counting. We calculated volumes of autofluorescence, MAdCAM-1, and clustered donor T cells (Volocity measurements: find objects using intensity > clip objects to ROI) within a 3D ROI fitted to border the whole PP. The lower limit of specific threshold intensity was calculated with ImageJ by the mean of 10 individual threshold measurements for each individual sample. For automated cell counting, we used the following algorithm: Volocity measurements: find objects using intensity > exclude objects touching edge of image > exclude objects by size > filter measurements: standard deviation. The lower limit of the object size was set to 400 μm2 as calculated before. We used a lower limit of the standard deviation ranging from 250 to 500 to ensure automated event identification. Mathematical analysis was performed with Excel (Microsoft) and statistical analysis with Prism 5 and InStat3 (GraphPad).
All animal studies were performed according to specific animal use protocols approved by Regierung von Unterfranken, State of Bavaria, Germany. All biopsy samples were obtained for diagnostic purposes and after written informed consent was obtained from donors. The study was approved by the Ethics Committee of the Medical Faculty, University of Würzburg, Würzburg, Germany.