Photoconversion of Alloreactive T Cells in Murine Peyer’s Patches During Acute Graft-Versus-Host Disease: Tracking the Homing Route of Highly Proliferative Cells In Vivo

The regulation of immune cell migration throughout the body is essential to warrant immunosurveillance and to maintain immune homeostasis. Marking and tracking of these cells has proven important to study mechanisms of immune cell trafficking and cell interaction in vivo. Photoconversion is a well-suited technique for intravital application because it enables contactless time- and location-specific marking of cells in the tissue without surgically manipulating the microenvironment of the cells in question. However, in dividing cells the converted fluorescent protein may decline quickly. Here, we provide a detailed description of the photoconversion technique and its applicability to tracking highly proliferating T cells from the priming site of T cell activation to peripheral target organs of effector function in a preclinical model. Dendra2+ T cells were photoconverted in the Peyer’s patches during the initiation phase of acute graft-versus-host disease (GvHD) and tracked through the mesenteric lymph nodes and the peripheral blood to the small intestine with flow cytometry and intravital two-photon microscopy. Photoconverted alloreactive T cells preserved the full proliferative capacity, homing, and migration of alloreactive T cells in the intestinal lamina propria. We conclusively proved that photoconversion of highly proliferative alloreactive T cells in the Peyer’s patches is an effective tool to study trafficking of alloreactive T cells under physiologic conditions and to GvHD target tissues. This technique can also be applied to the study of immune cell tracking under inflammatory and non-inflammatory conditions.

The regulation of immune cell migration throughout the body is essential to warrant immunosurveillance and to maintain immune homeostasis. Marking and tracking of these cells has proven important to study mechanisms of immune cell trafficking and cell interaction in vivo. Photoconversion is a well-suited technique for intravital application because it enables contactless time-and location-specific marking of cells in the tissue without surgically manipulating the microenvironment of the cells in question. However, in dividing cells the converted fluorescent protein may decline quickly. Here, we provide a detailed description of the photoconversion technique and its applicability to tracking highly proliferating T cells from the priming site of T cell activation to peripheral target organs of effector function in a preclinical model. Dendra2 + T cells were photoconverted in the Peyer's patches during the initiation phase of acute graft-versus-host disease (GvHD) and tracked through the mesenteric lymph nodes and the peripheral blood to the small intestine with flow cytometry and intravital two-photon microscopy. Photoconverted alloreactive T cells preserved the full proliferative capacity, homing, and migration of alloreactive T cells in the intestinal lamina propria. We conclusively proved that photoconversion of highly proliferative alloreactive T cells in the Peyer's patches is an effective tool to study trafficking of alloreactive T cells under physiologic conditions and to GvHD target tissues. This technique can also be applied to the study of immune cell tracking under inflammatory and non-inflammatory conditions. Keywords: t cell migration, acute graft-versus-host disease, mouse models, photoconversion, Dendra2, Peyer's patch, in vivo cell tracking, lymphocyte homing

INtroDUctIoN
Immune cells can migrate to distant locations within the body to warrant systemic immunosurveillance. This efficient immune cell homing is a prerequisite to protect against intruders or to regulate immune responses at any given location throughout the body. Specific immune surveillance becomes particularly impressive when comparing the volume of a T cell (35-95 fL) to the dimensions of the entire body and considering the small population size of a particular T cell clone with specific recognition capability (1). T cell homing plays an important role in many immunological reactions-therefore, delineating this process is central for understanding T-cell-mediated immunity. From a therapeutic perspective, it is essential to better understand the underlying mechanism of lymphocyte trafficking. This would allow specific blocking or fostering of homing routes in organ-specific inflammatory conditions such as autoimmune diseases, graft rejection after solid organ transplantation, or graft-versus-host disease (GvHD) following allogeneic hematopoietic cell transplantation.
Pioneering work in tracking of cell dynamics in vivo utilized transfer of radioactively labeled lymphocytes into rats, sheep, and other animals (2)(3)(4)(5)(6). Detecting radioactivity in different bodily fluids and organs indirectly proved the labeled cells' presence. Later, the discovery of congenic markers simplified transfer studies because it enabled the detection of adoptively transferred cells in vivo without needing to label the cells before transfer (7,8). The introduction of biocompatible fluorescent dyes permitted tracking of fluorescently labeled, adoptively transferred cells and their division cycle in vivo (9). When using cells expressing fluorescent proteins, they require no labeling before transfer or before detection analysis at all (10,11), and the fluorescence can even be induced at a specific time (12). Still, labeling dyes are widely used as a flexible means to label cells of interest before adoptive transfer (13).
However, none of these labeling methods are time-and site-specific at the same time, therefore it was not possible to mark specific populations of transferred cells after the transfer. Photoconversion is an excellent technique to mark cells in vivo in a specific location, because it enables contactless labeling without surgical manipulation of the organ of interest itself. This facilitates studying T cell homing without changing the homing properties of neither the studied cells nor the tissue surrounding the cells of interest. It is pivotal to minimize tissue perturbance and to avoid an experimental bias in the T cell homing process. By contrast, local introduction of a dye may lead to undesirable tissue perturbance and inflammation. Also, this would non-specifically label all cells present. Therefore, photoconversion is an expediant technique to introduce a time-and site-specific label specifically to transferred cells.
To date, several studies have employed photoconversion to monitor immune cell trafficking. The first (14) and most of the following studies employed the photoconvertible protein Kaede. These studies ranged from monitoring subcellular trafficking of single molecules (15) over organelles (16) to whole body trafficking of cell populations. Tracked cells comprised different precursor cells during embryogenesis (17,18), as well as immune cell populations (19) and pathogens (20). However, there are two limitations when transferring the findings of the mentioned studies to tracking T cells in vivo.
First, due to the development of new photoconvertible proteins, there are now even better choices for in vivo imaging. Due to its higher pK, the Dendra2 protein is photoconverted 20 times more efficiently than other convertible fluorescent proteins like Kaede and mEOS (21). This renders Dendra2 more suitable for low-illumination and deep-tissue photoconversion. Gurskaya and colleagues derived the photoconvertible fluorescent protein Dendra from the octocoral Dendronephthya sp (22). A subsequent A224V substitution in this protein resulted in Dendra2 with improved maturation and brighter fluorescence both before and after photoswitching (23). Photoconversion irreversibly switches Dendra2 green fluorescence from green to red. However, cell metabolism and cell division dilute the converted protein over time, especially in highly proliferating cells.
Second, immune cell trafficking was up to now studied mostly in non-proliferating cells or under steady state conditions (24)(25)(26). More limited are studies providing data from infection models (27,28) or cancer (29), where only a few specific T cell clones divide. Yet, at present, there are neither detailed reports on the photoconversion procedure nor systematic evaluation of possible side effects on T cell functions in vivo. Taken together, it remained unclear whether photoconversion would still be applicable to track an immune response with highly proliferative immune cells. Therefore, we tested whether we could employ photoconversion to track alloreactive T cells from the priming site to the peripheral target organ.
Here, we provide the first in-depth description of the intravital photoconversion procedure. We performed in vivo mouse experiments to study the homing from the Peyer's patches as priming sites to the mucosa of the small intestine as target site. We transplanted T cells expressing the photoconvertible protein Dendra2 in mitochondria (16). Subsequently, we photoconverted anatomical sites of initial T cell activation (30,31) just before trafficking starts, then tracked the cells for the subsequent 24 h to the target site.
We show that photoconverted alloreactive T cells can be tracked from the Peyer's patches through the mesenteric lymph nodes and the peripheral blood to the small intestines as target site, using flow cytometry. Furthermore, we demonstrate that photoconversion is well suited to observe T cell migration in target tissues with intravital microscopy.
Photoconversion does not impede cell division, homing, and migration in alloreactive T cells. Hence it is a suitable technique to study trafficking of highly proliferative cells in vivo. This protocol provides insight into the procedures and caveats of photoconverting alloreactive T cells in the Peyer's patches during acute GvHD. The photoconversion technique opens new opportunities for effective and exact tracking of cell distribution in vivo. It enables researchers to prove directly the time-and location-specific origin of a marked cell in any given tissue analyzed.

Materials
Materials are listed in Table 1 corresponding to the experimental procedures.

stEPWIsE ProcEDUrEs
This protocol focusses on the in situ photoconversion technique and subsequent flow cytometry analysis of converted donor T cells in distant sites. Therefore, these sections are in greater detail, whereas additional procedures are outlined in brief.
This study was carried out in accordance with the recommendations and regulations of the Regierung von Unterfranken.
The protocol was approved under the protocol number 55.2 2532-2-48.

Donor T Cell Enrichment
T cells were isolated from the spleen of Dendra2 + donor mice (16) at the age of 8-12 weeks. Splenocyte suspensions were enriched for T cells according to the manufacturer's protocol, counted by trypan blue exclusion, and adjusted to a density of 12 × 10 6 cells/ml in PBS. Typical T cell yields lay between 15 and 30% of the splenocyte input with a final T cell purity of 85-95%.

Donor Bone Marrow Isolation
Bone marrow cells were isolated from hind legs (femura and tibiae) of 8-to 12-week-old C57BL/6 mice. Cell numbers were determined by trypan blue exclusion, and the cell concentration was adjusted to 50 × 10 6 cells/ml. Typical bone marrow cell yields were 1-1.5 × 10 8 cells per mouse. Both the T cells and the bone marrow preparations can be stored at 4°C for up to 2 h until transplantation.

Transplantation
At the time point of transplantation, recipient mice should ideally have more than 20 g body weight to be sufficiently resilient to partial body weight loss. BALB/c mice at the age of 10-14 weeks were myeloablatively irradiated (8 Gy), and 1.2 × 10 6 Dendra2 + donor T cells were intravenously injected via the retro-orbital venous plexus together with 5 × 10 6 bone marrow cells in a total volume of 200 µl PBS ( Figure 1A). The drinking water was supplied with Baytril (Enrofloxacin, 0.05%) for 7 days after transplantation to avoid infections, and mice were monitored twice daily within this period. GvHD was scored clinically and body weight was measured daily.

Photoconversion
We photoconverted the cells 4 days after transplantation. This is the time point when the majority of donor cells are still in the secondary lymphoid organs and the first donor T cells leave the secondary lymphoid organs to enter the target tissues. This time point is optimal because short intervals between marking the cells and detecting them in the target tissue avoid dilution of the photoconverted protein before detection.

Preparation
Two sterile working spaces were prepared: one for animal preparation and one for the operation and photoconversion. For each of them, a heating mat was pre-warmed and covered with a sterile OP towel. All needed instruments and reagents were sterilized and laid out.
An operation cover was prepared by cutting an OP towel to a size of 15 cm × 15 cm, and an oval window was cut into the center with a diameter of 1.5 cm × 2.5 cm to be placed on the operation site.
A gauze swab was prepared by folding it to 7.5 cm × 5 cm and incising a 2 cm long central slit.
The collimator was connected to the UV lamp via the glass fiber and fixed on the laboratory stand.
A clean recovery cage was placed under an infrared lamp, and the sensor of a thermometer was placed on the bedding.

Operation and Photoconversion
Mice were anesthetized by injecting 10 µl anesthetic solution (Ketamine and Xylazine) per gram body weight intraperitoneally, and placed on the sterile, heated OP towel for preparation. Eyes were protected from dehydration by applying eye ointment. The ventral fur was wetted with 70% ethanol using a dissecting swab and shaved; the skin was disinfected with povidone iodine in an outward circling motion.
After 10 min, anesthetic depth was ensured to be stadium III.2 (surgical tolerance) by pinching the hind paw. If anesthesia was insufficient, further anesthetic was provided retro-orbitally. Not more than 20-50 µl was injected at a time to avoid over-dosage.
The prepared animal was transferred onto a fresh sterile and heated OP towel and the prepared OP-cover was applied. A 1-1.5 cm midline incision was made in the skin and in the peritoneum on the linea alba, no bleeding should be visible. The wound was covered with the prepared gauze swab, the slit overlaid on the incision. The swab was soaked with PBS to prevent drying of externalized intestinal tissue. Next, two cotton swabs were soaked with PBS and used to handle the intestinal contents. The swabs were inserted into the peritoneum to localize and gently exteriorize the cecum, which is located left craniolateral to the incision. The intestinal tissues were handled with great care to prevent a postoperative ileus. First, the cecum was placed on the gauze swab left lateral to the incision, and the intestinal tissue was kept moist at all times by dripping PBS on it using a Pasteur pipette. The small intestine was then gently pulled out, starting from the cecum, until the first Peyer's patch appeared ( Figure 1B). This was placed on the gauze swab facing upwards and covered with sterile tin film. A central hole in the foil with a diameter of 2 mm exposed the Peyer's patch but covered the surrounding intestinal tissue.
The collimator of the UV lamp was placed at a distance of 5 mm over the Peyer's patch, and the tissue was illuminated for 2 min at maximum power ( Figure 1C). An additional tinfoil shield protected the experimenters from scattered UV light, and UV protection glasses were used whenever the illumination was turned on.
The tin foil was removed, the intestine was gently pulled out until the next Peyer's patch appeared, and the procedure was repeated until six to seven patches were converted. In order to later be able to identify the converted patches during isolation, the distances in centimeter to the cecum and between the individual converted spots were documented for all converted patches.
After photoconversion, the small intestine and the cecum were gently placed back into the peritoneum as depicted in Figure 1D. Correct organization of intestinal loops was important to prevent a postoperative ileus. The peritoneum was closed with three non-consecutive stitches. Subsequently, the skin was closed with four stitches. The overall operation procedure took 30-45 min per animal. The animal was placed into a clean cage and held warm with an infrared lamp until recovered from anesthesia. The local temperature around the animal was monitored not to rise above 30°C to avoid overheating. When conscious, the animals were transferred back into their group cage. Analgesic was supplied in the drinking water.

Isolation of t cells
At different time points after photoconversion (0, 4, 12, and 24 h), mice were sacrificed, and the blood, mesenteric lymph nodes, Peyer's patches, and small intestine were harvested. One spleen was isolated for control stainings.

Peripheral Blood
Mice were anesthetized with Ketamine and Xylazin. After 10 min, the anesthetic depth was ensured to be stadium III.2 (surgical tolerance) by pinching the hind paw. The skin of the animals was disinfected, the animals were fixed on a pad and ventrally opened. The intestine was laid to the animal's left to expose the left renal vein and the vena cava inferior. An insulin syringe containing 200 µl of lysis buffer was inserted into the vena cava just below the fusion with the renal vein ( Figure S1 in Supplementary Material), with the opening facing cranial and ventral. Then, the needle opening was turned to the animal's left and gently pushed to the ventral side of the vena cava. Blood was collected very slowly to prevent collapsing of the vessel. Up to 700 µl were collected per animal. The erythrocytes were lysed in 30 ml of lysis buffer for 10 min at room temperature. The lysis was stopped by adding 20 ml of DPBS. The cells were pelleted for 5 min at 330 × g at 4°C and resuspended in 5 ml of PBS. All processed cell suspensions can be stored for up to 4 h at 4°C in DPBS until further analysis.
Afterward, the mouse was killed by cervical dislocation. The intestinal tract was removed from duodenum to rectum and placed on PBS-moistened paper towels. The mesenteric lymph nodes were excised, and all converted Peyer's patches were

Mesenteric Lymph Nodes
All surrounding tissue (fat and vessels) was carefully removed from the mesenteric lymph nodes. The lymph nodes were kept in a 1.5 ml tube containing 200 µl of PBS for up to 30 min on ice until further processing.

Peyer's Patches
Peyer's patches surrounded by intestinal tissue were spread out on PBS-moistened paper towels, and all surrounding tissue was precisely cut off using a scalpel. Exact trimming was fundamental for a good T cell recovery. The patches were kept in a 1.5 ml tube containing 200 µl of PBS for up to 30 min on ice until further processing. Mesenteric lymph nodes and Peyer's patches were thoroughly minced through a PBS-pre-wet 70 µm strainer using the sterile plunger of a 5 ml syringe, and rinsed through with 5 ml of icecold PBS. Cells were pelleted at 330 × g for 5 min at 4°C, and the supernatant was removed. Lymph node cells were resuspended in the residual volume and directly transferred completely onto a 96-well plate for staining. Cells from the Peyer's patch were washed once with 5 ml of ice-cold PBS, centrifuged, and then transferred onto the 96-well plate in the residual volume.

Intestine
Cells were isolated from the intestine using a protocol modified from Geem and colleagues (32). The small intestinal tube was opened longitudinally and vigorously dragged through 300 ml of PBS to remove fecal contents. It was cut into 2 cm segments and transferred into a 50 ml tube containing 30 ml of dissociation buffer. The tube was horizontally fixed on an orbital shaker set to 250 rpm for 20 min at 37°C. The solution was passed through a 100-µm strainer into a 50 ml tube. The incubation and filter step were repeated with another 30 ml of dissociation buffer. The first dissociation fraction was centrifuged at 330 × g for 5 min at 4°C and the pellet was washed in 30 ml of PBS. The two filtered cell suspensions were united and the PBS washing step was repeated. In initial experiments, the lamina propria fraction was additionally processed according to Geem et al. (32), but all donor T cells were found in the epithelial wash fraction ( Figure  S2 in Supplementary Material).

Spleen
A 70-µm cell strainer on a 50-ml Falcon was pre-wet with 2 ml of erythrocyte lysis buffer. The organ was placed on the strainer, incised crosswise a few times, minced with the plunger of a 5-ml syringe, and passed through the strainer by rinsing with 8 ml of erythrocyte lysis buffer. Erythrocytes lysis was stopped after 2 min by adding 10 ml of DPBS through the strainer. Cells were pelleted for 5 min at 330 × g at 4°C and resuspended in 5 ml of PBS.

Flow cytometry
Up to 1 × 10 6 cells were stained per well. Cells were resuspended in 100 µl blocking buffer and incubated for 5 min at 4°C. 100 µl of antibody mix was added, and cells were stained for 30 min at 4°C in the dark. Cells were pelleted at 330 × g for 5 min at 4°C and resuspended in 200 µl of PBS. The stained cells can be measured up to 6 h after staining when stored in the dark at 4°C. Longer pauses may require preceding fixation of the stained cells.
Fluorescence signal was acquired at 1,000-3,000 events/s, and washing wells were measured between different sample groups to exclude cross-contamination. The Dendra2 signal is extremely bright, especially in the FITC/GPF channel, and care was taken not to saturate any PMT. Table S1 in Supplementary Material contains a representative compensation matrix.
Donor T cells were identified by gating on lymphocytes, live cells, singlets, Dendra2-green, and/or Dendra2-red positive cells ( Figure S3 in Supplementary Material).

Proliferation Assay
Splenocytes were isolated from Dendra2 + mice as described in Section Spleen. Half of the cells were photoconverted in a black 24-well dish for 2 min with a rotating motion of the glass fiber and without the collimator. 5 × 10 6 cells/ml were stained with 1 µl of cell trace violet for 6 min at room temperature. 500 µl of FCS and 5 ml of RPMI were added successively, and samples were incubated at 37°C in a water bath for 5 min. The samples were pelleted at 330 × g for 5 min at room temperature and resuspended in 5 ml of PBS. Cells were seeded on a 96-well round bottom plate at 200,000 cells/well in 200 µl of proliferation medium and left to proliferate for 3-4 days.

In Vivo cell Migration
Mice were anesthetized and an intestinal loop was exposed as described in the photoconversion operation procedure (see Operation and Photoconversion). Instead of wetted gauze and OP towels, cling foil was used to avoid lint sources in the microscope. The mouse was positioned on a heating pad and the intestinal loop was positioned under a glass cover slip using two custommade holders. The tissue was kept moist using sterile 0.9% NaCl solution. Fluorophores were excited at a wavelength of 840 nm. The light intensity was increased as the square of penetration depth along the z-axis, between 5 and 30%. Images were acquired every 30 s in a field of view of 495 µm × 495 µm with a resolution of 256 × 256 pixels and a distance of 3 µm between two z-planes.

cytotoxicity Assay
Lymphocytes were isolated from the axillary, brachial, cervical, inguinal, and mesenteric LN and the spleen of naïve Dendra2 + mice. The containing T cells were stimulated in a mixed lymphocyte reaction (MLR) with allogenic BALB/c splenocytes that had been irradiated with 30 Gy or that had been T-cell-depleted. 8 × 10 6 B6; Dendra2 + responder cells were incubated with 4 × 10 6 BALB/c stimulator cells in 2 ml/well of RPMI with 10% FCS, 1% Pen/Strep, and 1 mM sodium pyruvate in 24-well plates for 4-5 days at 37°C, 5% CO2. Cytotoxicity was assessed in round bottom 96-well plates with 50,000 Luciferase-expressing MOPC-315.BMP2.FUGLW target cells (33) and 50,000 to 5 × 10 6 MLR effector cells containing cytotoxic T lymphocytes. Target cells only were used as a negative control, 0.1% Triton was used as a positive control. Survival of the target cells was quantified from the bioluminescence signal. The cells were pelleted at 330 × g for 5 min and resuspended in 110 µl of 2.75 mg/ml Luciferin in PBS. The cells were incubated for 3 min at room temperature and the signal was measured in a black 96-well plate. Percent of lysis was calculated with the formula (experimental signal − background signal)/(maximal signal − background signal) × 100.  stencil with a diameter of 2-3 mm (Figure 2A). The in situ photoconversion efficiently switches the majority of Dendra2 + donor cells in the Peyer's patch from green to red. About 90% (range: 75-98%) of donor cells isolated immediately after photoconversion contained photoconverted Dendra2 protein ( Figure 2B). We detected photoconverted cells only in the Peyer's patches, and no photoconverted cells leaked into the blood or the small intestine at this time point (Figure 2C). Only single cells resided in the mesenteric lymph nodes immediately after the photoconversion procedure, which took up to 30 min per mouse. In conclusion, photoconversion of the Peyer's patches reliably marks the majority of currently residing Dendra2 + donor T cells in the converted tissue area.

t cells retain their Proliferative capacity
To address whether photoconversion of T cells would affect key biological functions such as T cell proliferation, we compared the proliferative capacity of photoconverted T cells versus untreated T cells in vitro. Both populations proliferated to the same extent (Figures 3A,B). To find out how long the photoconverted Dendra2-red remains detectable in rapidly dividing cells, we measured the decay of red fluorescent signal during proliferation. Stimulated splenocytes lost the photoconverted Dendra2-red signal at the same rate as the proliferation dye, with half of the red signal being retained after each division cycle. The photoconverted population separated from the untreated control for at least four divisions before the two populations merged (Figure 3C). In this setting, CD8 T cells had completed two to five or more cycles and proliferated more vigorously as compared to the CD4 T cells, which had divided one to four times after 4 days proliferation ( Figure 3D). We measured the activation marker CD44 before and after stimulation on converted versus unconverted T cells. Photoconversion did not affect the expression of CD44 with or without stimulation after 4 days in culture ( Figure 3E). Overall, photoconversion does not affect the cells' ability to proliferate. Notably, non-dividing photoconverted Dendra2-red + T cells retained their fluorescence intensity at least 4 days in cell culture.

t cells traffic From Peyer's Patches to the Mesenteric lymph Nodes, blood, and Intestine
To evaluate whether photoconversion would affect T cell motility or efficient homing, we analyzed the homing behavior of photoconverted Dendra2 + donor T cells at different time points after photoconversion (Figure 4). After 4 h, first T cells had left the Peyer's patches and resided in the mesenteric lymph nodes, where 0.5% of donor lymphocytes originated from the photoconverted Peyer's patches. In the peripheral blood, only 0.04% of donor lymphocytes originated from the Peyer's patches. 12 h after photoconversion, cell numbers had increased in both sites, mesenteric lymph nodes and peripheral blood, as compared to 4 h after photoconversion, and reached even higher numbers after 24 h. Notably, the vast majority of donor T cells in the Peyer's patch still remained positive for the red fluorescent signal at this time point. We readily detected T cells in the small intestine after 24 h at a frequency of 4.3% (mean). Photoconversion Does Not Impair t cell Migration Within the target organ or cytotoxicity After entering the target organs by extravasation, alloreactive T cells migrate in the tissue to find their target cells and mediate the potentially lethal graft-versus-host response. We tested whether the photoconverted T cells performed equally well in migrating in the target organ, and found that the photoconverted cells were able to migrate as well as the green cells in the lamina propria of the small intestine ( Figure 5A). Migration speed as a basic measure of the ability to migrate, and the distribution of turning angles as a measure of directionality were not affected by photoconversion (Figures 5B,C). Furthermore, we assessed whether photoconversion affects the cells' cytotoxicity. The photo converted cells performed similarly in lysing the target cells in vitro ( Figure 5D).
These results indicate that homing, migration, and effector functions of alloreactive T cells are preserved after photoconversion. We can track the cells on their way from the priming to the target site without perturbing the course of the immune reaction.
In summary, photoconverted T cells fully retained the ability to proliferate, to traffic to distant sites, and to migrate within tissues at a time point between the initiation and effector phase of acute GvHD. Therefore, we conclude that photoconversion is suitable to study adaptive immune responses such as the

DIscUssIoN
Our findings demonstrate that Dendra2-expressing T cells can be efficiently photoconverted in the Peyer's patches during the induction phase of acute GvHD. Photoconversion yields proliferating cells that traffic via their regular homing route to the intestine. There, they migrate within the tissue to find their target antigen. Hence, photoconversion proved to be applicable to this highly proliferative cell population for at least four cell divisions.
The main advantages of this technique are • Contactless and efficient labeling of cells.
• Effective and exact tracking of cells in vivo.
• Direct evidence about time and location of origin in any given organ analyzed.
We photoconverted the cells toward the end of the initiation phase. This is the time when the first cells start entering the small intestinal lamina propria, and therefore, it is a suitable time point to track the early effector phase in the target organ. At the time point of photoconversion, more than 90% of cells in the Peyer's patches express CD44 and the majority have divided (34). We readily detected the first infiltrating cells in the lamina propria 24 h after photoconversion, which corresponds to day +5 after transplantation. At this time point, the donor cells had a larger size than when transplanted as measured by flow cytometry (data not shown), indicating a T cell blast phenotype. Since at comparable time points, almost all donor T cells proliferated (35), one can expect that more than 90% of the cells detected have undergone multiple divisions. This means that low numbers of proliferating cells can be detected with high specificity. Care has to be taken to exclude cell debris and dead cells in order to receive results that are biologically relevant (see also Figure S3 in Supplementary Material). Although there is a contamination of other cell types in the transplanted enriched T cell population (5-15%) which will also be photolabeled if present in the Peyer's patches, the vast expansion of the donor T cells in an allogenic setting outgrows the contamination by far. Furthermore, contaminating non-immune cells will most likely not home to the host lymphoid organs after transplantation. Therefore, we expect that all donor cells detected 5 days after transplantation are indeed donor T lymphocytes. The majority of these cells is still positive for the photoconverted fluorophore at the time point of detection 24 h after photoconversion. Therefore, later time points during the effector phase are likely to yield similarly positive results, since more cells will leave the Peyer's patches during the course of disease.
It is not possible to determine the maximum time after conversion that converted cells remain detectable in the intestine. This is because the mice die from GvHD about 4 days after photoconversion (about 8 days after transplantation) in this model. Nevertheless cells proliferate greatly in GvHD. It is thus likely that the photoconversion signal would be diluted out after five division cycles in these cells (Figure 3). This is expected to be 24-48 h after photoconversion assuming a division rate of 6 h (36). This is distinctly different from the lack of dilution of photoconverted Dendra2 protein in a non-proliferative setting. In such steady state circumstances, it is possible to follow the fate of photoconverted cells for several months (28). A further problem arises if the cells to be converted are in large or highly pigmented organs. The converting light may then not penetrate deeply enough. Hence, it is necessary to ensure efficient penetration by the converting light before performing homing analyses to grant complete conversion of the residing cells. Furthermore, highly pigmented organs heat up more than normal tissues under the UV illumination. In such organs, careful measures are needed to avoid tissue overheating and drying. Extended troubleshooting guidelines are displayed in Table 2. More factors may have to be adjusted when using other photoconvertible proteins which photoconvert less efficiently. Finally, there were no visible alterations of the microenvironment, but we cannot exclude minor phototoxicity to cell populations that were not relevant to this study. In brief, limitations of this technique are • Limited accessibility of very thick or highly pigmented organs.
• Potential phototoxicity affecting cell populations that were not analyzed in this study.
Notably, we found that even immediately after photoconversion, single converted cells can be detected in the mesenteric lymph nodes. This was surprising given that the two organs are far apart. We postulate that these cells may have resided in the efferent lymphatics of one of the first converted Peyer's patches, and had traveled to the mesenteric lymph nodes within the 30-40 min of photoconversion and blood collection. Such rapid migration is entirely realistic assuming a lymph flow rate of 100 µm/s (38).
Our photoconversion protocol generated a high percentage of converted cells (90%). This is considerably higher than when the intestinal lymphocytes were converted endoscopically (30%) or where cells in cervical lymph nodes were converted through the skin (46%) (24). The numbers are comparable to another study converting Dendra2 in the Peyer's patches (99%). In contrast to our study, their protocol included much shorter but more intense ultraviolet light exposure (27).
We conclude that photoconversion of highly proliferative alloreactive T cells in the Peyer's patches as presented in this protocol is a suitable and effective tool for studying physiological trafficking of T cells during inflammatory disease conditions such as acute GvHD. This protocol can be directly applied to the study of T cells under physiological steady state settings, in cancer models or under infectious conditions, to investigate cell migration. With some adjustments, it can also be applied to other cell types or to other animal models.

EtHIcs stAtEMENt
This study was carried out in accordance with the recommendations and regulations of the Regierung von Unterfranken. The protocol was approved under the protocol number 55.2 2532-2-48.

AUtHor coNtrIbUtIoNs
KJ designed, planned, and carried out the experiments, analyzed and interpreted the data, and wrote the manuscript. ZM analyzed and interpreted the cell migration tracks. LS, JH, and ST helped in the photoconversion and cell isolation. D-DL, HS, MR, and MQ helped in cell isolation. KH helped with the establishment of ultraviolet illumination and gave feedback on the project. AB designed the study, interpreted the data, and revised the manuscript. All authors read and discussed the manuscript.

AcKNoWlEDGMENts
We thank Tim Steinfatt and Dr. Julia Delgado-Tascón for advice with the cell isolation and stimulation, and Mike Friedrich and Jürgen Pinnecker for help with the photometer and the two-photon microscopy setup. We thank Dr. Niklas Beyersdorf for advice on the mixed lymphocyte reactions. We would like to express our gratitude to Drs. Andreas Brandl, Julia Delgado-Tascón, Michael Sixt, Manfred Lutz, Eleni Koutsilieri, Andrew Davis, and the whole Beilhack lab for valuable advice and discussion.
FIGUrE s1 | Scheme displaying the collection of blood from the vena cava. After ventral opening of the mouse, the intestines were laid to the right and the vena cava was exposed. A 30 G syringe containing 200 µl of lysis buffer was inserted at the junction of the left renal vein with the inferior vena cava and blood is slowly collected from the vessel.