In vivo magnetic resonance imaging of dendritic cell migration into the draining lymph nodes of mice

Authors


Abstract

Dendritic cell (DC) migration into the draining lymph nodes is critical for T cell priming. Here, we show that magnetic resonance imaging (MRI) can be used to visualize DC migration in vivo. We combined clinically approved small particles of iron oxide (SPIO) with protamine sulfate to achieve efficient uptake by murine bone marrow-derived DC. SPIO-DC were largely unaltered and after injection into the footpads of mice, they migrated into the T cell areas of the draining lymph nodes, which could be visualized by MRI. Distinct MRI signal reduction patterns correlated with the detection of SPIO-DC mainly within Thy-1.2+ B220 T cell areas, as confirmed by iron staining and immunohistology. Clear signal reduction patterns could still be observed with 1 × 106 injected SPIO-DC at high resolution, resulting in the detection of about 2000 DC. Control injections of homing-incompetent SPIO-DC derived from CCR7–/– mice or SPIO alone did not reach the T cell areas. Taken together, the results demonstrate that clinically approved contrast agents allow the non-invasive visualization of DC migration into the draining lymph node by MRI in vivo at high resolution. This protocol therefore also allows dynamic imaging of immune responses and MRI-based tracking of human DC in patients.

Abbreviations:
MRI:

magnetic resonance imaging

SPIO:

superparamagnetic iron oxide particles or small particles of iron oxide

Introduction

DC are the major cell type to sense pathogens and initiate adaptive immune responses in vivo1. Therefore, DC are an attractive tool for therapeutic manipulation of the immune system and are currently tested for the treatment of cancer 25. Although the first clinical trials could demonstrate the potential of DC therapy, the efficiency of the vaccination needs further improvement. Factors influencing DC immunogenicity depend on the subset, maturation state, activation stimulus, route, dosage and frequency of DC injection 3. Their migration from the site of infection to the draining lymph node represents a major requirement for their immunogenic function and depends on the expression of the chemokine receptor CCR7 as a lymph node homing receptor 6, 7. After s.c. injection less than 1% of injected DC reach the lymph nodes, and this can be enhanced up to tenfold by preconditioning of the injection site with DC or proinflammatory cytokines such as TNF, while CCR7-deficient DC still fail to migrate 7. Therefore, effective migration of the injected DC to the secondary lymphoid organs remains an essential step for DC therapy 8 and efficient monitoring of DC migration by a non-invasive method is required.

The biodistribution of [111In]-labeled tumor-peptide-pulsed DC injected intradermally or intranodally into melanoma patients could be imaged by gamma scintigraphy, but without any spatial resolution and anatomical correlation 9. Due to its high resolution and the possibility of non-invasive imaging (MRI) has recently gained attention for tracking of magnetically labeled cells in vivo, including neuronal and mesenchymal stem cells or lymphocytes. For cellular MRI, the cells need to be labeled with a contrast agent in order to be differentiable from the surrounding tissue. For this purpose, many studies have used superparamagnetic iron oxide particles (small particles of iron oxide, SPIO) for efficient labeling of different cell types 10. SPIO are completely biodegradable and clinically approved. Their phagocytic uptake is mediated by scavenger receptor SR-A, and larger particles with a diameter of 120–180 nm (USA: Feridex®, Europe: Endorem®) show higher uptake rates than smaller ones (20–40 nm) 11. Due to their size, intravenously injected SPIO are rapidly taken up by the reticulo-endothelial system (RES, i.e. liver Kupffer cells) and are therefore routinely used for liver imaging in the clinic 12, 13. One study used receptor-mediated endocytosis of anti-CD11c antibody-conjugated SPIO by DC. However, the resolution for the detection of 1× 107 SPIO-DC at the injection site was very low and no migration to the draining lymph node could be detected 14.

By combining SPIO with commonly used transfection agents, molecular complexes are formed through electrostatic interactions, which enhance cellular uptake 1517. Stem cells and other mammalian cells could be efficiently labeled by the combination of SPIO with protamine sulfate, which is clinically used to reverse heparin anti-coagulation, but also in vitro as a cationic transfection agent 18, 19.

Here, we combined SPIO with protamine sulfate to achieve highly efficient DC labeling. After s.c. injection of 1 × 106 DC there was still a clear MRI signal reduction detectable in the T cell areas of the draining lymph node. The MRI data correlated with the immunohistological detection of SPIO-DC in these lymph nodes. Control injections of SPIO-DC derived from CCR7-deficient mice or SPIO alone revealed no or only subcapsular signal reductions, respectively. Taken together, our data demonstrate that highly efficient SPIO-labeling of DC with clinically approved reagents enables the non-invasive visualization of their migration in vivo by MRI.

Results

BM-DC labeling with SPIO and protamine sulfate

For in vitro and in vivo MRI experiments with SPIO- DC it was important to determine the optimal conditions for effective labeling of cells. Protamine sulfate has been shown to increase the uptake of SPIO by cells at different concentrations 18. Titration experiments were performed in which BM-DC were incubated with 5 µg protamine sulfate and SPIO at different iron concentrations ranging from 0 to 400 µg Fe/mL. MRI measurements revealed the strongest signal reduction by cells, which had been incubated with 100 µg Fe/mL and 5 µg/mL protamine sulfate. At higher iron concentrations (200 µg/mL, 400 µg/mL), signal reduction was decreased (Fig. 1A). Similar results were seen when Prussian blue staining was performed on cytospins of labeled DC to detect incorporated SPIO (data not shown). To define the MRI detection threshold of labeled DC in vitro, titration experiments with different numbers of labeled cells were performed. The lowest detectable signal reduction could be determined at 1 × 104 SPIO-DC/well incubated at 37°C in the presence of protamine sulfate, corresponding to a resolution of 7.33 cells/voxel (Fig. 1B). In contrast, DC labeled with SPIO in the absence of protamine sulfate displayed a decreased signal reduction intensity and a detection threshold of 3× 104 cells per well, corresponding to a resolution of 21.98 cells/voxel.

Figure 1.

Efficient labeling of DC with SPIO and protamine sulfate (A) BM-DC were incubated with SPIO at different iron concentrations (0, 50, 100, 200, 400 µg Fe/mL) and protamine sulfate (5 µg/mL) for 4 h at 4°C or 37°C. Cells were washed three times with 10 U heparin/mL PBS, pelleted in 1. 5mL tubes and measured by MRI. The intensity of signal reduction can be correlated with the iron content per cell. (B) BM-DC were incubated with SPIO at 100 µg Fe/mL with or without 5 µg/mL protamine sulfate for 4 h at 4°C or 37°C. After three washing steps with 10 U heparin/mL PBS, labeled cells were titrated with unlabeled cells, resuspended in 1% agarose, immediately placed in the wells of a RIA plate, and measured by MRI. Each well contains 3 × 105 cells with a decreasing ratio of labeled to unlabeled cells (3.0–0.03 × 105, top to bottom). Three similar experiments have been performed with comparable results. (C–E) BM-DC were incubated with SPIO (100 µg Fe/mL) and protamine sulfate (5 µg/mL) for 4 h at 4°C or 37°C, followed by three washing steps with 10 U Heparin/mL PBS and stimulation with TNF over night. (C) Cytospins were prepared and cells were stained with Prussian blue (dark grey) and counterstained with Nuclear Fast Red (light grey). (D) Cells were analyzed by transmission electron microscopy. (E) Magnifications of D. All experiments were performed at least two-to-three times with similar results.

Internalized SPIO are directed to endosomal compartments

To confirm that signal reduction observed in the in vitro MRI experiments was caused by internalized SPIO and not by particles sticking to the outer membrane, iron detection was performed with SPIO-DC on cytospins and by electron microscopy. Prussian blue staining revealed iron incorporation only in cells incubated at 37°C, indicating that not surface adherence but active cellular uptake of SPIO occurs in DC (Fig. 1C). To finally prove the incorporation of SPIO by BM-DC, transmission electron microscopy was performed. Dense iron-containing endosomal compartments were present in DC incubated at 37°C, but not when incubated at 4°C (Fig. 1D, E). Moreover, SPIO-labeling remained stable within DC for at least 4 days in vitro (Supplementary. Fig. 1).

Migration and T cell priming of SPIO-DC in vivo

To assess the question, whether the SPIO labeling procedure and the deposited iron may cause any changes in the immunological properties of DC, different functional assays were performed in vitro and in vivo. FACS analysis revealed no significant changes in the surface expression of molecules involved in antigen capture and presentation (CD205, MHC-II), costimulation (CD40, CD80, CD86), or adhesion and migration (CD54, CCR7) between unlabeled and SPIO-labeled DC (data not shown). Furthermore, labeling of BM-DC with SPIO did not alter DC viability as assessed by Annexin V-binding assays (data not shown). To elucidate the impact of SPIO-labeling on the migratory capacity of DC in vivo, unlabeled (DC) and SPIO-labeled (DC+SPIO) matured DC were additionally stained with CFSE (Fig. 2A). CFSE+ cells could be correlated with MHC IIhigh expression and iron incorporation (Fig. 2A). Unlabeled and SPIO-labeled CFSE+ DC were injected into TNF-presensitized footpads of mice, as preconditioning of the injection site with proinflammatory cytokines leads to an increased DC migration into the draining lymph nodes 7. CFSE+ cells localized in Thy1.2+ B220 T cell areas of the draining popliteal lymph nodes after 24 h (Fig. 2B). FACS-analysis revealed similar levels of immigrated unlabeled and SPIO-labeled CFSE+ DC into the popliteal lymph nodes with a maximum after 24 h (Fig. 2C) while the total lymph node cellularity further increased (Supplementary Fig. 2). After 72 h, the absolute number of immigrated SPIO-DC was reduced in comparison with unlabeled DC (Fig. 2C). To investigate the life span of injected DC versus SPIO-DC both populations were labeled with CFSE and after lymph node removal counter-stained with propidium iodide. The results indicated comparable survival rates of both populations (not shown). To further characterize the functional quality of SPIO-DC to prime T cells, in vivo priming experiments were carried out. While priming of allogeneic T cells in the popliteal lymph nodes after s.c. injection was comparable between unlabeled and SPIO-DC (Fig. 2D), the capacity to prime syngeneic T cells with KLH protein was slightly reduced (Fig. 2E). These findings may indicate that the in vivo antigen presentation capacity remains intact but SPIO-loading may have minor influences on migration and the uptake or processing of antigens.

Figure 2.

Influence of SPIO labeling of DC on their migratory and T cell priming capacity in vivo. For quantification of in vivo migrated DC, day 8 BM-DC were labeled with SPIO, matured, and co-labeled with CFSE. (A) Cytospins were prepared, DC were stained against surface MHC II and iron was detected by Prussian blue staining. (B) After injection of 5 × 106 CFSE-SPIO-DC into TNF-presensitized skin they accumulate in the T cell areas as detected by counterstaining for Thy1.2 and not in B cell areas (B220). (C) SPIO-labeled or unlabeled DC (1 × 106) were additionally labeled with CFSE and injected into TNF-presensitized right or left footpads of mice, respectively. After 24, 48 and 72 h, popliteal lymph nodes were analyzed by flow cytometry. Percentages of CFSE+ propidium iodide DC per injected cell number within the lymph node are indicated. (D) SPIO-labeled or unlabeled DC (1 × 106) derived from BALB/c mice were injected into the TNF-presensitized right and left footpads of allogeneic C57BL/6 mice. After 10 days, popliteal lymph nodes of DC-injected and uninjected control mice were removed and restimulated with titrated numbers of mitomycin C-treated BALB/c spleen cells. Proliferation was measured after 3 days. (E) SPIO-labeled or unlabeled DC (1 × 106) derived from BALB/c mice were loaded with KLH protein and injected into TNF-presensitized right and left footpads of syngeneic mice. After 10 days, popliteal lymph nodes of DC-injected and uninjected control mice were removed and restimulated with titrated amounts of KLH protein. Proliferation was measured after 3 days. Experiments were carried out in triplicates for each condition and bars represent the means ± SD. Two similar experiments for each type of analysis have been performed with comparable results.

Lymph node homing of SPIO-DC followed by MRI

Because the functional assays described above did not show strong reductions in immunological key features of DC by labeling procedures, SPIO-DC were used for in situ MRI. To increase the migration rate, mature SPIO-DC (1 × 107/foot) were injected twice at days 0 and 1 s.c. into mouse footpads and high resolution MRI scans of popliteal lymph node regions were performed on day 2. The control lymph node with immigrated unlabeled DC appeared homogenously grey (Fig. 3, open arrow). In contrast, a strong signal reduction could be observed in the lymph node where SPIO-DC had been injected s.c. (Fig. 3, closed arrow).

Figure 3.

SPIO-labeled DC home to the draining lymph node. BM-DC (BALB/c) were labeled with SPIO or remained unlabeled and 1 × 107 matured cells were injected into the right (R) or left (L) foot pad of a syngeneic mouse, respectively, followed by a second injection of cells 24 h later. Another 24 h later the mouse was sacrificed and high resolution MRI of popliteal lymph node regions was performed (scan time 13:30 h). A series of six 500-µm thick axial slices ranging from proximal to distal is shown in (A–F). The right lymph node (R) shows strong signal reduction (filled arrow) where SPIO-DC were injected into the foot pad. In contrast, no signal reduction (open arrow) can be observed in the left lymph node (L) where unlabeled DC were injected. More than 15 similar experiments have been performed with comparable results.

MRI signals correlate with the number of injected SPIO-DC

To determine the lowest number of injected SPIO-labeled DC that can be detected in the draining lymph node by MRI, cell numbers of injected DC were titrated in consecutive experiments. A clear signal reduction was observed in defined areas of the lymph node even at low numbers of 1 × 106 injected SPIO-DC (Fig. 4A–D, left column). For detection of lower cell numbers, TNFα-presensitization preceded the DC application. As shown in Fig. 5C and D, a single injection of 5 × 106 or 1 × 106 SPIO-DC resulted in a distinct signal reduction pattern in the central parts of the draining lymph nodes, which implied that the distribution of labeled DC could be responsible for this T cell area-like pattern. Control lymph nodes showed no comparable signal reduction and appeared homogenously grey (Fig. 4A–D, right column).

Figure 4.

SPIO-DC can be detected dose-dependently in the central T cell areas of the draining lymph node. DC (BALB/c) were labeled with SPIO, matured and the indicated cell numbers were injected into TNF-presensitized skin or DC were injected twice at days 0 and 1 into the right foot pads of BALB/c mice. High resolution MRI was performed on day 2 and a series of six 300–500 µm adjacent slices of popliteal lymph node regions are shown for each experiment. Scan times were 13:30 h (A), 6:50 h (B), 6:24 h (C), 10:15 h (D). Arrows indicate the images with the strongest signal reduction. Results could be reproduced in at least two independent experiments for each condition. Size bar: 2 mm. (E) To visualize DC-homing to the central T cell areas of the draining lymph nodes, immunohistochemistry was performed on serial sections of dissected lymph nodes. Representative stainings are shown for a popliteal lymph node 24 h after injection of 5 × 106 SPIO DC into the presensitized foot pad. T = T cell area, B = B cell area.

Figure 5.

Injected wild type SPIO-DC, but not SPIO-DC from CCR7–/– or soluble SPIO alone, reach the central T cell areas of the draining lymph nodes. (A, B) CCR7–/– SPIO-DC were injected into TNF-presensitized skin or injected twice into the hind food pads of mice. At day 2, mice were sacrificed and popliteal lymph node regions were measured by MRI at high resolution. (C, D) SPIO alone were injected in 50 µL PBS (100 µg/mL) into the right foot pads of mice. Mice were sacrificed after 4 or 24 h and measured at high resolution by MRI. A series of six 300 µm adjacent axial slices of popliteal lymph node regions are shown for each experiment (A–D). Scan times were 12:50 h (A), 6:50 h (B), 11:00 h (C), 11:30 h (D). Scale bar: 2 mm. Two similar experiments for each type of analysis have been performed with comparable results. (E) Matured wild type or CCR7–/– SPIO-DC (1 × 106), SPIO in PBS, or PBS alone were injected s.c. into TNF-presensitized food pads of mice. At day 2, mice were sacrificed and popliteal lymph node regions were measured by MRI at high resolution over night. Scale bar: 1 mm. Photographs of the same dissected lymph nodes show no iron coloration from the outside after injection of CCR7–/– SPIO-DC. While PBS-injected mice revealed small lymph nodes without signal reduction, the lymph nodes draining the skin of mice injected with SPIO in PBS showed a subcapsular signal reduction in the SPIO-lymph node correlated with a corona-shaped iron coloration on the surface of the dissected lymph nodes (arrow). (F) Prussian blue staining of the sectioned lymph nodes after CCR7–/– SPIO-DC or SPIO alone injection confirm the signal reduction measured by MRI. Four similar experiments for each type of analysis have been performed with comparable results.

Histological analysis of the dissected lymph nodes confirmed that the signal reduction observed by MRI was indeed caused by SPIO-labeled DC. In addition, a combination of Prussian blue staining and immunohistochemistry allowed the visualization of SPIO-DC colocalizing mostly within central Thy1.2+ B220 T cell areas (Fig. 4E), which is similar to the distribution patterns of CFSE+ SPIO-DC shown in Fig. 2B. In contrast, no iron could be detected in the control lymph nodes (data not shown).

Taken together, these experiments allowed the detection of about 2000 injected SPIO-DC in the lymph node by MRI at injected cell numbers of 1 × 106 DC.

SPIO-DC migration depends on CCR7

To further elucidate the correlation observed between MRI signal reduction and distribution of labeled DC in popliteal lymph nodes, two experimental settings were performed as additional controls. First, mature CCR7–/– SPIO-DC were injected according to the experiments described above. Mice were scanned by MRI, lymph nodes were dissected afterwards and analyzed by histology. Due to their lack of the lymph node homing receptor CCR7 6, 7, CCR7–/– DC are not able to migrate into lymph nodes, which should result in a different pattern detected by MRI. Indeed, as shown in Fig. 5A, B, E and F, signal reduction could not be observed in the central parts of the draining lymph nodes after CCR7–/– DC injection. However, MRI scans revealed a slight signal reduction in distinct parts of the subcapsular region that appeared to be of a dose-dependent nature and decreased with lower numbers of injected cells (Fig. 5A, B). These observations could be further supported by histology of dissected lymph nodes, as Prussian blue staining revealed few iron deposits only in the subcapsular sinus regions at high cell numbers injected. (Fig. 5F).

In a second set of control experiments, soluble SPIO were injected directly into footpads and their distribution in the draining popliteal lymph nodes was measured by MRI. High resolution MRI scans revealed a corona-like shaped signal reduction pattern after 4 and 24 h (Fig. 5C–E). The corona-like shape of iron oxide distribution could also be seen macroscopically under the surface of dissected lymph nodes (Fig. 5E, arrow) and was further confirmed by Prussian blue staining of lymph node sections (Fig. 5F). These findings are consistent with published data, that high-molecular weight particles like Endorem® (diameter 120–180 nm) should be drained to the sinus areas of the lymph node, but not enter deeper regions 20.

These findings further support our theory that only fully functional migratory SPIO-DC, but not passive transport of SPIO, is responsible for the signal reduction in the central T cell areas of draining lymph nodes observed by MRI.

SPIO-DC detection by MRI in vivo

To this point, all mouse MRI measurements had been performed in situ of sacrificed mice. Next, we examined the distribution on SPIO-labeled DC in vivo and followed the kinetics of DC migration by measuring popliteal lymph nodes at different time points after injection. For this purpose, 1 × 106 mature SPIO-DC or unlabeled control DC were injected into presensitized right and left footpads of mice, respectively, and MRI measurements of the draining lymph node regions were performed under anesthesia at different time points prior and post DC-injection. Because of relatively short in vivo MRI measurements, the resolution was lower than in the in situ measurements described above. No signal reduction could be observed in the lymph node before and 4 h after injection (Fig. 6, arrow and Supplementary Fig. 3). In contrast, at 16 h post injection, increased signal reduction patterns appeared in streaks proceeding from the periphery into the more central parts of the lymph node. By 24 h, the signal reduction intensity had further increased, and by 48 h, was more concentrated in the central T cell areas (Fig. 6, arrows). A remarkable increase in lymph node size was also obvious in MRI scans and dissected lymph nodes. From 72 h onwards post injection, the signal reduction patterns appeared more scattered and the signal reduction intensity began to decrease (Fig. 6, arrows). No similar signal reduction patterns could be detected in the control lymph node, although the increase in lymph node size was comparable with that of the SPIO-DC draining lymph node (Fig. 6, right column). Correlating the clear signal reduction observed at 24 h here (Fig. 6) with the absolute number of DC detected at the same time point by CFSE labeling and FACS analysis (Supplementary Fig. 2), the detection of our in vivo measurements can be assumed as 2000 SPIO-DC within a lymph node.

Figure 6.

Dynamic in vivo MRI of SPIO-DC migration into popliteal lymph nodes confirms in situ MRI data. Mature SPIO-DC or unlabeled DC (1 × 106) were injected into TNF-presensitized right or left footpads of one mouse, respectively. The 30–45-min MRI scans of the draining popliteal lymph node regions were performed under isoflurane anesthesia at indicated time points prior and post injection of DC. A series of six 500 µm adjacent axial slices is shown for the right popliteal lymph node region for each measurement. In addition, one representative axial slice of the left control lymph node region is also shown. Scan times were 30–45 min. The data is representative for three independent experiments.

Discussion

The aim of this study was to establish an efficient labeling method for BM-DC that allows their non-invasive in vivo tracking by MRI.

The determination of the most effective iron concentration for labeling of BM-DC led to an optimal ratio of 100 µg Fe/mL SPIO to 5 µg/mL protamine sulfate, which is consistent with previously published data 18. Protamine sulfate, an antidote to heparin anticoagulation used in the clinic 21, 22 was also shown to facilitate the uptake of SPIO by cells 18. Protamine sulfate is also used as a transfection agent in molecular biology and is well tolerated by cells 19. It forms complexes with SPIO through electrostatic interactions, resulting in an efficient uptake by human mesenchymal stem cells (MSC), hemopoietic (CD34+) stem cells and other mammalian cells 18. Remarkably, iron incorporation by BM-DC was not only reduced at lower concentrations of incubated SPIO, but was also reduced at higher SPIO concentrations, as shown in the titration experiments evaluated by MRI and Prussian blue staining of cytospins. As noted by Arbab et al. 18, this might be caused by a disequilibrium of electrostatic interaction between the carboxyl groups on the dextran-coated SPIO and the polycation protamine sulfate, as microscopically visible particles could be observed, depending on the ratio of SPIO to protamine sulfate.

Further elucidation of the labeling efficiency led to the determination of the MRI detection threshold of labeled cells in vitro. The data presented in this study revealed an in vitro MRI resolution of 7.33 cells/voxel for DC incubated with SPIO and protamine sulfate, representing 1 × 104 labeled cells/100µL. Furthermore, the SPIO-uptake-facilitating effect of protamine sulfate could be demonstrated, as DC labeled without protamine sulfate displayed a decreased in vitro MRI resolution of 21.98 cells/voxel. Due to different experimental settings and measurement protocols, these numbers are difficult to compare with the results of other studies. Several aspects have to be considered, for example the use of different cell types and transfection agents for labeling procedures. In the study by Kircher et al. 23, the detection threshold for tat-CLIO-labeled CD8+ T cells was calculated to be <2 cells/voxel in vitro (3 × 104 cells/50µL) and 3 cells/voxel in vivo. In a different study, in which hematopoietic progenitor cells were labeled with SPIO and P7228 liposomes, a detection threshold of 2.5 × 105 cells has been reported 24. Matuszewski et al. 25 calculated a detection threshold of 1 × 104 labeled tumor cells, measured on a clinically used 1.5-T magnet.

Compared to other studies, the labeling procedure applied here shows substantial advantages, especially with respect to a possible implementation of SPIO-DC-based cell tracking for the clinic. In contrast, although the HIV-derived Tat peptide conjugates with iron particles resulted in efficient labeling of various cell types, including stem cells 26 or T cells 23, 27, the use of HIV-derived products in clinical settings remains unclear. All agents we used for the DC labeling procedure are clinically approved, i.e. SPIO (Endorem®), protamine sulfate, and heparin, which is used in the washing steps. Furthermore, SPIO-labeling did not impair phenotypical and functional characteristics of the DC, as we could not observe changes in surface marker expression, apoptosis induction, migration and T cell priming in vivo. This is consistent with SPIO-labeling of other cell types 18, 25, 28.

For an improved homing of DC, presensitization of the injection site with DC or TNF results in an up to tenfold increase in the rate of migrated DC in the draining lymph nodes 7. The initial dose comprising two consecutive injections of 1 × 107 DC could be reduced to TNF-presensitization and one injection of 1 × 106 cells per footpad, which were sufficient to observe distinct migration patterns by means of MRI. Titration of injected cell numbers led to the detection of distinct signal reduction patterns mainly within the central T cell areas of the lymph nodes. Overlapping distribution patterns of immigrated SPIO-DC within T cell areas could be confirmed by Prussian blue staining and immunohistology on cryosections of dissected lymph nodes.

Control experiments showed that CCR7–/– SPIO-DC do not cause a signal reduction in the central areas of the lymph node and SPIO alone do only cause a subcapsular staining. This is consistent with other reports, in which CCR7–/– DC did not reach the central areas of the lymph nodes 6, 7, and soluble high-MW molecules, like SPIO, did not enter the central paracortex areas 20. Concerning the dose-dependent detection of few subcapsular SPIO deposits in CCR7–/– SPIO-DC-injected draining lymph nodes, it is difficult to judge whether this is due to migrated CCR7–/– DC, migration of endogenous DC which have captured dying CCR7–/– DC, or fluid phase transport of SPIO alone which could have been released from dead cells. Taken together, the negative controls support the hypothesis that signal reduction patterns detected in distinct regions of the draining lymph nodes are indeed caused by immigrated SPIO-DC within T cell areas.

Since the in situ MRI measurements had shown a correlation of signal reduction and immigrated SPIO-labeled DC for as low as 1 × 106 injected cells, the approach was adapted for tracking DC in vivo. Due to shorter measuring times on anesthetized mice, it was not possible to measure in high resolution. However, distinct signal reduction patterns could be observed over time, and corresponded with the in situ MRI data. After 4 h, the lymph nodes displayed no signal reduction patterns in the central regions and the lymph node size was comparable to the wild type before injection of cells. This is consistent with previous studies in which fluorescent dye-labeled DC had not yet migrated into the T cell areas by 4 h, but at 8 h and later time points post injection 29. In line with these findings, MRI measurements after 16, 24 and 48 h revealed signal reduction patterns in the central areas of the lymph node, comparable to the in situ MRI measurements. After injection of 1 × 106 CFSE-labeled SPIO-DC, about 0.2% of them could be detected after 24 h by FACS in the draining lymph node. Thus, a clear MRI pattern at 24 h correlates with an absolute number of about 2000 CFSE+ DC measured by FACS per lymph node. Furthermore, MRI measurements displayed an increase in lymph node size, similar to observations by Martin-Fontecha et al. 7, in which injection of DC resulted in an increase in lymph node cellularity that was detectable on day 1 and reached a plateau by days 3 and 4. By 72 h after injection, signal reduction patterns began to diminish, which may be explained by dilution effects due to the increasing lymph node size. Apoptosis of SPIO-DC may also be considered as it has been reported by others 30. However, our in vitro data indicate that DC survive well for 4 days and counterstaining of CFSE+ DC with propidium iodide reveal similar survival rates also in vivo. Detection of migrated DC by CFSE label reached a maximum after 24 h and declined thereafter, while MRI measurements show maximum signal reduction after 48 h. The decline of detectable CFSE+ DC within the lymph node could reflect the lower sensitivity of the CFSE method in our hands or due to quenching of the tracer. Our MRI measurements are, however, consistent with the literature 7. The SPIO-CFSE-DC appear to decline more rapidly at 72 h as compared to control CFSE-DC. Since apoptosis does not seem to apply, only the impaired migration at later (72 h) but not early (24 h) time points could be considered. This may also be reflected by the slightly reduced KLH priming. The reasons for differences between early and late migration need further investigation.

In vivo experimental approaches to study DC migration are complex and different technologies are available. Most of these experiments involve adoptive transfer of labeled DC. However, visualization of the transferred cells by microscopic techniques still requires intervention, which is not applicable for human use. For intravital two-photon imaging, for example, lymph nodes have to be surgically prepared in order to gain close access for the objective. Although this technique allows dynamic imaging of interactions between DC and T cells in anesthetized animals, there are some other limitations, including the imaging depth of maximally 500 µm 31. Experiments with radiolabeled DC could provide non-invasive insights into the migratory pathways of injected DC. For example, i.v. injected DC accumulate in the lung and are then redistributed to the spleen and liver or intradermally administered DC rapidly home to the T cell areas of the lymph node. However, these measurements could only image the distribution two-dimensionally, without any spatial resolution and anatomical correlation 9, 3234. The nuclear medicine techniques positron emission tomography (PET) and single photon emission computed tomography (SPECT) also utilize radiopharmaceutical agents to track cells in vivo35, 36. However, like gamma scintigraphy and bioluminescence imaging 37, PET lacks spatial resolution, a prerequisite for the precise anatomical localization of migrated DC in vivo.

In this study, we concentrated on the MRI technology, as it provides a superior anatomical correlation and spatial resolution. In comparison with two recently published studies, which also used MRI as a means of tracking labeled DC in vivo, our labeling approach appears to result in a higher resolution in situ and in vivo, although we used a weaker magnet 38, 39. In the study by de Vries et al. 39, a different migration pathway was investigated. There, MRI was used to visualize SPIO-labeled DC injected at high numbers (7.5 × 106) directly into the lymph nodes of melanoma-patients and to monitor DC migration into adjacent lymph nodes 39. We followed DC migration from the footpad into the draining lymph node and could detect clear DC distribution patterns in the T cell compartments with only 1 × 106 injected DC, or subcapsular signal reduction when only SPIO were injected. The study by Ahrens et al. 38, shows an elegant labeling-approach allowing the discrimination between perfluoropolether-labeled DC and the surrounding anatomy. We tracked immigrated SPIO-DC in the lymph node for several days in vivo and were able to generate a kinetic view of DC migration patterns together with the increase of total cellularity of the draining lymph node by MRI.

Taken together, this study demonstrates that efficient labeling of DC with clinically approved reagents allows the non-invasive visualization of their migration into the draining lymph nodes in vivo by MRI. This technique may therefore provide further insights into the dynamics and kinetics of DC migration in vivo and could be applied for tracking of systemically administered DC in animal models and humans.

Materials and methods

Mice

BALB/c and C57BL/6 mice were obtained from Charles River (Sulzfeld, Germany) and were bred in the animal facilities of the Department of Dermatology. CCR7–/– (BALB/c background) mice were kindly provided by Martin Lipp (Max-Delbrück-Center for Molecular Medicine, Berlin, Germany) and obtained from Klaus Schröppel (Institute of Clinical Microbiology, Immunology and Hygiene). All animal experiments were carried out according to German laws and guidelines.

Generation of murine BM-DC

The preparation and culture of bone marrow cells from BALB/c and C57BL/6 mice to generate DC has been described in detail before 40. Briefly, 2 × 106 BM cells were cultured for 8–10 days with 10% supernatant from a GM-CSF-producing cell line 41. DC maturation was performed overnight with 500 U/mL TNF (Peprotech/Tebu, Frankfurt, Germany).

Labeling of BM-DC with SPIO

BM-DC were cultured until day 8–10, harvested, washed, counted and adjusted to 1 × 107/mL serum-free RPMI medium (Biowhittaker, Cambrex Vervier, Belgium) and 10% GM-CSF supernatant. SPIO (Endorem®, 11.2 mg Fe/mL, Guerbet, Aulnay-sous-Bois, France) and protamine sulfate (10 mg/mL, American Pharmaceutical Partner, Schaumburg, USA) were prediluted in serum-free RPMI plus 10% GM-CSF sup. Of this solution, 100 µL was added in titrated doses or, if not otherwise indicated, 1:1 to the cell suspensions resulting in final concentrations of 100 µg Fe/mL and 5 µg/mL, respectively, and incubated for 4 h at 37°C. DC were washed three times with 10 U heparin/mL PBS and matured with TNF overnight in RPMI with 10% FCS and 10% GM-CSF sup. The next day, cells were washed in PBS and adjusted to appropriate cell numbers for the indicated experiments.

Prussian blue staining and immunohistology

For the preparation of cytospins, 2 × 104 cells in 200 µL medium were spun down on microscope slides with a Cytospin 3 centrifuge (Shandon, Astmoor, UK) for 3 min at 500 rpm and room temperature. Prussian blue staining was used to detect SPIO. Samples were fixed with 4% paraformaldehyde for 30 min at room temperature, washed, and then incubated for 30 min with 2% potassium ferrocyanide in 2% hydrochloric acid, washed again with aqua dest., and counterstained with Nuclear fast red. Another washing step with aqua dest. was followed by dehydration in ethyl alcohol (90, 95, and 100%) and cover-slipping.

To analyze the in situ distribution of SPIO-labeled DC, draining popliteal lymph nodes were dissected 24 h after administration of DC, embedded in OCT compound (Tissue-Tek®), and snap-frozen in liquid nitrogen. The 5–10-µm-thick cryosections were prepared on microscope slides, air dried and frozen at –20°C until staining procedures.

Immunofluorescence stainings with antibodies against MHC II (TIB-229 supernatant, ATCC), Thy1.2 (53–2.1, BD PharMingen), and B220 (RA3–6B2, BD PharMingen), or the appropriate isotype controls, were performed on fixed samples for 30 min at room temperature, followed by Alexa Fluor® 555-conjugated secondary antibodies (Molecular Probes).

For immunohistochemistry, fixed lymph node sections were incubated with protein blocking solution (DakoCytomation), followed by incubation with primary antibodies against Thy1.2 and B220, and appropriate biotinylated secondary antibodies. After extensive washing in TBS, sections were incubated with StreptABComplex alkaline phoshatase (DakoCytomation) for 30 min and washed again. Sections were developed with Fuchsin (DakoCytomation) according to manufacturer's instructions. Levamesole (Sigma) was added to reduce endogenous alkaline phosphatase activity. Subsequent Prussian blue staining was performed as described above, except counterstaining. Images were acquired with a LEICA DMRD microscope operated by the Improvision Openlab™ software and processed with Photoshop software (Adobe Systems).

In vivo quantification of migrated SPIO-DC

DC were labeled with 1 µM CFSE (Molecular Probes) for 10 min at 37°C, washed extensively, and then injected s.c. into the hind footpads of mice. After 24 h, popliteal lymph nodes were removed, single-cell suspensions were prepared, and measured by FACS analysis (FACScan, Becton Dickinson). For the evaluation of SPIO-DC migration into draining lymph nodes by immunofluorescence microscopy, DC were alternatively labeled with 5 µM CFSE and used as described above.

Electron microscopy

Cells were incubated with SPIO at 37°C or 4°C as described above. Further handling of the samples was kindly performed by Andrea Hilpert at the Department of Anatomy I. In brief, cells were fixed in 2.5% glutaraldehyde in PBS (pH 7.4), pelleted, post-fixed in 1% osmium tetroxide, dehydrated and embedded in Epon resin. Ultrathin sections were cut using a Leica UCT microtome and stained with lead citrate. Images were acquired by transmission electron microscopy on a Zeiss EM 906 equipped with a plan film camera.

In vivo priming

BM-DC (BALB/c) were loaded or not with SPIO and matured with TNF overnight. DC (1 × 106) were injected s.c. into the hind foot pads of allogeneic C57BL/6 mice. After 10 days, popliteal lymph nodes (4 × 105 cells/well) were restimulated in a 96-well flat-bottom plate (Becton Dickinson) at titrated numbers with mitomycin C-treated spleen cells from BALB/c mice. After 3 days, the triplicate cultures were pulsed with 1 µCi [3H]thymidine (Amersham) for 16 h and harvested onto filtermats with an ICH-110 harvester (Inotech, Dottikon, Switzerland); filters were counted with a 1450 Microplate Counter (Wallac, Turku, Finland).

For KLH priming, the same protocol was applied with the only difference that syngeneic DC were used and co-incubated also with 50 µg/mL KLH (Calbiochem/Merck, Darmstadt, Germany) together with the TNF-maturation stimulus. For restimulation, the draining lymph node cells (4 × 105 cells/well) were exposed to titrated amounts of KLH as indicated.

MRI measurements

MRI was performed on a 4.7 T BRUKER Biospec scanner (Ettlingen, Germany) with a free bore of 40 cm, equipped with an actively RF-decoupled coil system. For determination of the most effective iron concentration for DC-labeling in vitro titration experiments were conducted by analyzing DC pellets in 1.5 mL tubes without agarose. In vitro titration of labeled cells was performed in wells of custom-fit RIA plates. The cells in agarose have a dimension of 5 mm diameter and 6,4 mm height. MRI slice thickness was 1 mm. The whole slice was positioned inside of homogenous part of cell pellet content in such a way that influence of surface irregularities and air was neglectable. A quadrature mouse volume coil from Rapid Biomedical (Wuerzburg, Germany) was used as a receiver and transmitter coil for the in situ/in vivo experiments. The scanning procedure started with the acquisition of T2 weighted spin echo coronal anatomical images for localization of lymph nodes (slice thickness 1 mm, 16 slices, field of view 35 x 35 mm, matrix 256 x 128, TR = 2800 ms, TEef = 77 ms, 1 average) using a rapid acquisition relaxation enhanced sequence (RARE) with rare factor equal 8. In situ axial images were acquired using multi slice multi echo (MSME) sequence (slice thickness 0.35 mm, 10 slices, field of view 30 x 30 mm, matrix 256 x 256, TR = 3000 ms, TE = 10 ms, number of echoes 16, 30–50 averages). In vivo axial images were acquired using MSME sequence (slice thickness 0.35–0.5 mm, 10 slices, field of view 30 x 30 mm, matrix 256 x 128, TR = 3000 ms, number of echoes 16, TE = 10 ms , 6 averages). In vivo images were processed with Photoshop software (Adobe Systems). Individual lymph nodes were identified manually and assigned to new layers. Then, a threshold of 90 was set for these layers, resulting in further darkening of low-intensity (i.e. SPIO) areas, while higher-intensity (i.e. lymph node tissue) areas remained unchanged.

Acknowledgements

This work was supported by the Interdisciplinary Center for Clinical Research (IZKF), projects B7 for M.B.L. and Z2 core unit for A.H. and L.B., and the Doerenkamp Professorship for Innovations in animal and consumer protection for K.B. We thank Martin Lipp and Klaus Schröppel for providing/breeding the CCR7–/– mice, Susanne Rößner and Jens Hänig for their help with the CFSE-DC injections, Andrea Hilpert for the help with the electron microscopy, and the histology section of the Dermatology Department for their help with the Prussian blue staining.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

Ancillary