Dynamic Imaging of T Cell-Dendritic Cell Interactions in Lymph Nodes Sabine Stoll, Jérôme Delon, Tilmann M. Brotz, and Ronald N. Germain

Cell preparation and transfer: Bone marrow cells from B10.A mice were cultured for 7 days in GM-CSF (5 ng/ml) and IL-4 (10 ng/ml) (1). Cells recovered from these cultures were generally 80 - 90% CD11c⁺. The CD11c⁺ BMDC were MHC class II^med-hi, CD80^med, CD86^low-med, CD40^neg. These BMDC were either left antigen-free or pulsed for 12 hrs with 3 M PCC 88-104 peptide in the original culture, then labeled for 15 minutes at 37°C with 5 M DiI, DiD, or SNARF, washed, and 1.5 x 10⁶ labeled cells injected subcutaneously into the hind footpad of normal B10.A mice. The dyes used were found in pilot experiments not to interfere with DC migration into the draining LN (2). These same studies showed that labeled DC began to appear in the draining LN in appreciable numbers approximately 6 hrs. after injection (2). CD4⁺ T-cells were obtained from the spleen and LN of unmanipulated 5C.C7 TCR transgenic (3) x RAG-2^-/- (4) mice (Taconic Farms, Germantown, NY). T-cells were labeled with 5 M CFSE in all experiments except that shown in Fig. 4, in which 5 M SNARF was used. DiI and DiD appeared to inhibit T-cell migration under these conditions (2) and were not used with T cells for this reason. Thirty million labeled T-cells in 250 l were injected i.v. into B10.A mice that had previously received BMDC in a modification of the method of Ingulli et al. (5). All mice were housed and manipulated in accord with approved procedures.

Lymph node preparation and imaging: The draining popliteal lymph node was removed at the indicated time after T-cell i.v. transfer, which followed the s.c. injection of antigen-free or antigen-bearing DC by 18-37 hrs. as indicated in each figure or supplemental movie. Preliminary experiments confirmed that the transferred T-cells homed to the expected locations within lymph nodes (Fig. S1), but unexpectedly revealed that simultaneous administration of T-cells and antigen-bearing, but not antigen-free, DC retarded arrival of the DC in the lymph node (2). For this reason, DC were given 6-12 hrs. prior to the T-cells in all the experiments reported here. A lymph node containing transferred T-cells and DC was trimmed free of fat, placed in a heated chamber (delta T dish, Bioptechs, Inc., Butler, PA), and fixed in place with partially cooled 2.5% agarose. The dish was then placed into the stage adapter on the microscope, filled with 200 l Iscove's Modified Dulbecco's medium without phenol red containing 10% fetal calf serum, 35 mM HEPES, 2 mM glutamine, penicillin / streptomycin, 1 mM sodium pyruvate, non-essential amino acids, and 50 M 2-mercaptoethanol and covered with a heated lid to maintain the system at 37°C. Following immobilization of the lymph node in the Bioptechs chamber, imaging was performed using a Leica TCS-SP2 confocal microscope and a 40x oil immersion objective. The specimen was illuminated using 488 nm argon, 568 nm krypton, and / or 633 nm helium lasers, depending on the dyes used in the particular experiments. Detector slits were configured to minimize any cross talk between the channels. 512 x 512 pixel density images were collected every three minutes as a z-stack of 16 images, each 5 m thick. Printed images show a maximum projection of the z-stack at the indicated time point.

Cell migration tracking: Individual T-cells were tracked over a 3 hr interval, plotting their position every 3 min. when a new image was acquired. The mean distance traversed was determined from these plotted migration paths, the known image magnification, and a reticule scale. This was divided by the total time of the recording, yielding mean velocity.

Preparation and imaging of naïve T-cells expressing CD43-GFP: 5C.C7 T-cells activated and then infected in vitro as reported (6) did not show appreciable accumulation in the draining lymph node, as expected from the loss of CD62L and, presumably, CCR7 expression. Therefore, a variation of the method of Hollander et al. (7) was employed to generate naïve CD43-GFP⁺ T-cells that would home to LN. 5C.C7 x RAG-2^-/- TCR transgenic mice were treated with 150 mg/kg 5-fluoruracil to mobilize bone marrow stem cells. Three days after drug treatment, bone marrow was harvested from the long bones and subjected to 6 rounds of spin infection using MSCV-CD43-GFP retrovirus as previously described for in vitro T-cell infection (6, 7). The infected bone marrow cells were transferred i.v. into sublethally irradiated (700 rad) B10.A mice. After 7 weeks, LN and spleen cells were harvested to provide a lymphocyte population consisting in large measure of CD43-GFP⁺ CD4⁺ 5C.C7 T-cells. These cells were labeled with SNARF and injected into B10.A mice that had been injected s.c. in the footpad 12 hrs previously with 1.5x10⁶ BMDC pulsed for 12 hrs with 3 M PCC 88-104 peptide. Lymph nodes were harvested after 23 hrs and imaged as described above, except that the pixel density was increased to 1024 x 1024 from 512 x 512 and the images were collected with a z-step distance of 1 m. The resulting data sets were processed using the Autoaligner and Imaris 3.1.2 softwares (Bitplane, Zürich, Switzerland) to provide both x-y and x-z images, the latter corresponding to an en face view of the putative immunological synapse between the T-cell and BMDC. The Surpass option of Imaris was used to generate a space-filled rendering of the T-DC interaction.

Supplemental Figure 1. Frozen sections of lymph node were prepared 18 hrs after injection of CFSE-labeled T-cells (green) and SNARF-labeled BMDC (red) into syngeneic animals. The sections were stained for CD19 (blue) (A) or for CD3 (blue) (B). T-cells and the few DC seen in the sections were only found in the T-zones and not the B-cell follicles.

Medium version | Full size version

Supplemental Figure 2. CD69 expression by CFSE⁺ T-cells recovered from imaged lymph nodes similar to those employed in Text Figure 1A and B and containing antigen-unpulsed (blue open histogram) or antigen-pulsed (red closed histogram) DC.

Medium version | Full size version

To view these movies, download a QuickTime viewer.

• Movie S1A   size = 1,474KB

• Movie S1B   size = 1,985KB

  T-cells establish stable contacts with antigen-bearing DC. DiI-labeled antigen-bearing DC (red) were injected subcutaneously. Six hrs later CFSE-labeled 5C.C7 CD4⁺ TCR transgenic T-cells (green) were injected intravenously and popliteal lymph nodes were prepared 18h later. A z-stack of images consisting of 16 optical sections of 5 m thickness was collected every 3 min. over a period of 15 hrs. Each movie represents the maximum projection images produced from the z-stacks. (A) shows all data points collected in the last 5.7 hrs of a 15 hr experiment, while (B) represents the full 15 hr collection with only every third z-stack projection shown. Text Figure 1 contains images from this data set.

• Movie S2   size = 1,375KB

  Higher magnification of Movie S1. T-cells remain in close association with DC, staying with a DC even when the APC migrates.

• Movie S3   size = 1,217KB

  DiI labeled antigen-bearing DC (red) were injected subcutaneously 12 hrs prior to intravenous inoculation of CFSE labeled 5C.C7 CD4⁺ TCR transgenic T-cells (green). Twenty hrs later popliteal lymph nodes were prepared and imaged as for Movie S1, except that the total time of data collection represented in the movie is 3.8 hrs.

• Movie S4   size = 2,338KB

  Visualization of T-cell migration in a lymph node. DiI-labeled antigen-bearing DC (red) were injected subcutaneously. Six hrs later CFSE-labeled 5C.C7 CD4⁺ TCR transgenic T-cells (green) were injected intravenously and popliteal lymph nodes were prepared 37 hrs later, then imaged as for Movie S1, except that the total time of data collection represented in the movie is 6.5 hrs.

• Movie S5   size = 2,235KB

  Higher magnification of Movie S4. T-cells migrate rapidly throughout the lymph node.

• Movie S6  size = 1,607KB

  Visualization of T-cell migration in a lymph node. DiI-labeled antigen-bearing DC (red) were injected subcutaneously. Six hrs later CFSE-labeled 5C.C7 CD4⁺ TCR transgenic T-cells (green) were injected intravenously and popliteal lymph nodes were prepared 32 hrs later, then imaged as for Movie S1, except that the total time of data collection represented in the movie is 4.2 hrs.

• Movie S7

  Observation of T-cell division in a lymph node. DiI-labeled antigen-bearing DC (red) were injected subcutaneously. Six hrs later CFSE-labeled 5C.C7 CD4⁺ TCR transgenic T-cells (green) were injected intravenously and popliteal lymph nodes were prepared 37 hrs later, then imaged as for Movie S1. The cell division shown here was imaged for 5.8 hrs starting 10 hrs after data collection began.

Note S1: The number of BMDC in the lymph node studied is similar to that found in the presence of an infectious / dangerous stimulus. In previous work, as many as 10⁴ additional CD11c⁺ cells were found in the lymph node draining the site of skin bombardment with gold particles using a gene gun (8), which is an accepted vaccination strategy. Similar increases have been seen using systemic LPS or STAg inflammatory stimuli (9, 10). Others have shown that the number of Langerhans cells that migrate to a lymph node after skin painting with a contact sensitizer is similar to the number of BMDC arriving in the lymph nodes examined here and these LC have a similar surface phenotype to the BMDC at the time of transfer (MHC class II⁺⁺, CD40-, CD80⁺, CD86^low). The number of DC transferred has also been varied and no differences from the results reported seen with as few as 0.3x10⁶ injected DC.

Note S2: While single photon imaging causes more photodamage than multiphoton imaging under otherwise similar conditions, such cell damage is unlikely to be the origin of the cell behavior reported here. With LN removed both early and late after T-cell transfer, the imaging involved the same instruments, laser settings, and time of exposure. Thus, phototoxicity should be similar in both cases, even though cell migration was much greater in the latter situation. In addition, no apparent decrement in cell mobility was seen after several hours of imaging with LN taken more than 37 hrs. after T-cell transfer into animals given containing antigen-bearing DC.

Note S3: Experiments were also conducted in which just T-cells but no DC were transferred. Under these conditions, no movement of the T-cells is seen at early or late times after transfer, in contrast to the results with antigen-bearing DC. This makes it very unlikely that the BMDC desensitize the T-cells by secreting excess chemokine, given that there is a gain in migratory activity following introduction of these cells with bound antigen. Finally, BMDC make high levels of SLC and ELC, but this is also true for DC activated in vivo and secretion of these mediators is key to attracting T-cells to DC for scanning of displayed antigen. Whether even higher levels are achieved using the transferred cells as compared to inflammatory stimuli applied in situ is unknown. In any case, excess chemokine signaling is unlikely to have accounted for the lack of T-cell migration in lymph nodes removed at early times after T-cell transfer. Migration was seen only when antigen was present, although chemokine production should have been the same with pulsed versus unpulsed BMDC. In addition, high levels of SLC overcome the stop signal imposed by antigen recognition (11), so the effect of excess chemokine production should be greater not less migration in the presence of antigen, the opposite of what was observed.

Direct measurements have shown lymphoid tissues to have low oxygen tension under in vivo physiological conditions (12); it is actually of a concern that in room air, oxygenation may be greater than normal, especially at the more superficial areas of the lymph node. The first widely employed in vitro system for generating immune responses (13) used a gas mixture with 7%O₂ and the reducing agent 2-ME is typically added to lymphocyte culture medium when this work is performed at normal 20% O₂. T-cells move well when recordings involve lymph nodes removed late after cell transfer as compared to early, even though both have the same conditions of oxygenation; furthermore, experiments using SNARF, a pH sensitive dye, do not show a change in pH within the T cells during the recording periods employed here. It is therefore unclear whether attempts to enhance oxygenation are in fact more or less likely to give conditions that best reflect the in vivo state. The obvious solution is to do the microscopy under intravital conditions with normal blood flow. However, it is extraordinarily difficult to make recordings at high resolution over prolonged times with exteriorized lymph nodes of living, breathing animals. Thus, it will be necessary to combine judicious use of short-term experiments under intravital conditions with long-term tracking studies in an explant model such as the one employed here so that a more complete view of lymphocyte behavior can be achieved.

Note S4: Exclusion of CD43 can be also occasionally be observed in the region of contact of T-cells with transferred unpulsed DC and even with what appear to be host DC. Therefore, it seems that the exclusion of CD43 is not strictly foreign-antigen dependent. However, T-cells make productive interactions with self-MHC molecules in vivo, as shown with respect to maintenance of partial TCR phosphorylation (14, 15) and even various aspects of immunological synapse formation (16), so such molecular polarization is not necessarily unexpected. In any case, these data do not distinguish TCR-mediated CD43 exclusion from exclusion due to chemokine signaling. The latter has been suggested by observations of CD43 polarization at the rear of T cells migrating in a chemokine gradient in vitro (17). A distinction between TCR and chemokine receptor-driven exclusion of CD43 from the contact region cannot be made using the present transfer system because of the potential capacity of the (invisible) endogenous DC to interact with the T-cells through their self-MHC ligands. This experiment will require use of a different transgenic model (AND or OT-II), which select in mice with the H-2^b haplotype and which would permit use of A^-/- mice that lack class II expression on endogenous cells, into which I-A^b expressing MHC class II⁺ DC can be transferred with or without antigen pre-exposure.

Note S5: Although the lymph node preparations imaged here lack blood and lymph flow, as well as signals from nerves, the time at which CD69 elevation is seen, at which cell division is seen, and at which dissociation and movement of T-cells away from antigen bearing DC is seen matches results obtained in completely in vivo models. CSFE labeled naïve TCR transgenic T-cells first show evidence of cell division at approximately 36 hrs after antigen exposure begins in vivo (18); likewise, CSFE-labeled cells are sequestered from the blood stream for 48 hrs in the presence of antigen.

References

1. K. Inaba et al., J. Exp. Med. 176, 1693 (1992).

2. S. Stoll, unpublished observations.

3. R. A. Seder, W. E. Paul, Annu. Rev. Immunol. 12, 635 (1994).

4. Y. Shinkai et al., Cell 68, 855 (1992).

5. E. Ingulli, A. Mondino, A. Khoruts, M. K. Jenkins, J. Exp. Med. 185, 2133 (1997).

6. J. Delon, K. Kaibuchi, R. N. Germain, Immunity 15, 691 (2001).

7. G. A. Hollander, B. D. Luskey, D. A. Williams, S. J. Burakoff, J. Immunol. 149, 438 (1992).

8. A. Porgador et al., J. Exp. Med. 188, 1075 (1998).

9. C. Reis e Sousa et al., J. Exp. Med. 186, 1819 (1997).

10. C. Reis e Sousa, R. N. Germain, J. Immunol. 162, 6552 (1999).

11. S. K. Bromley, D. A. Peterson, M. D. Gunn, M. L. Dustin, J. Immunol. 165, 15 (2000).

12. C. C. Caldwell et al., J. Immunol. 167, 6140 (2001).

13. R. I. Mishell, R. W. Dutton, J. Exp. Med. 126, 423 (1967).

14. J. R. Dorfman, I. Stefanova, K. Yasutomo, R. N. Germain, Nature Immunol. 1, 329 (2000).

15. D. Witherden et al., J. Exp. Med. 191, 355 (2000).

16. P. Révy, M. Sospedra, B. Barbour, A. Trautmann, Nature Immunol. 2, 925 (2001).

17. J. M. Serrador et al., Blood 91, 4632 (1998).

18. C. Tanchot, D. L. Barber, L. Chiodetti, R. H. Schwartz, J. Immunol. 167, 2030 (2001).