A direct comparison of rejection by CD8 and CD4 T cells in a transgenic model of allotransplantation

Given the high frequency of HA-specific CD8⁺ T lymphocytes known to circulate in transgene-positive (Tg⁺) Clone 4 mice [18], we hypothesized that HA104 skin grafts would be rapidly rejected by Tg⁺ Clone 4 hosts in an antigen-specific fashion. We observed consistent and vigorous rejection of HA104 skin grafts by Tg⁺ Clone 4 mice (n=9, MST: 14.2 days; Fig. 1). Rejection correlated with a dense mononuclear cellular infiltrate consistent with acute cellular rejection (data not shown). Transgene-negative Clone 4 littermates (Thy 1.1⁺) accepted HA104 grafts (Thy 1.2⁺) indefinitely, indicating that rejection is dependent on the expression of the transgene-encoded TCR (n=3, MST>100 days, p=0.006 vs. transgene-positive animals; Fig. 1). Additionally, HA-negative BALB/c skin transplanted to Tg⁺ Clone 4 mice was accepted indefinitely (n=4, data not shown). Collectively, these results indicate that both transgenic Clone 4 T cells and graft expression of HA are required for rejection in this model.

To exclude the possibility that CD4⁺ T cells expressing the transgenic TCR in the Clone 4 mouse were also required for allograft rejection in this system, we eliminated CD4⁺ lymphocytes from Clone 4 hosts prior to transplant using the depleting anti-CD4⁺ antibody GK1.5. CD4-depleted hosts remained capable of rejecting HA⁺ skin grafts, indicating that CD8⁺ cells are capable of independently mediating rejection in transgenepositive Clone 4 hosts (2/4 reject, MST: 26.5). That rejection by depleted animals is somewhat less consistent and occurs with a delayed tempo suggests that CD4⁺ T cells may enhance the response and that this model may allow future exploration of subset cooperativity.

Collectively, these results indicate that CD8⁺ T cells from Clone 4 hosts consistently reject HA104 skin grafts. Rejection is dependent on expression of the trangene-encoded TCR and HA expression in the graft. That CD4-depleted Clone 4 mice remain capable of rejecting skin grafts suggests that CD8⁺ T cells alone are capable of mediating rejection.

To confirm that graft rejection by the Clone 4 transgenic mouse mimics cellular rejection in normal non-transgenic hosts, we evaluated whether Clone 4 hosts develop a memory response to HA antigen. When Tg⁺ Clone 4 mice that had previously rejected HA104 skin were regrafted with HA104 skin, graft rejection recurred with a significantly accelerated tempo, consistent with an anamnestic response (n=7, MST: 8 days, p=0.008 vs. naïve Clone 4 hosts). We subsequently demonstrated that this memory response could be transferred into syngeneic BALB/c hosts who do not reject HA104 skin. Following complete rejection of HA104 skin grafts, 5×10⁵ lymphocytes from Tg⁺ Clone 4 hosts were isolated and injected via tail vein into BALB/c hosts grafted with HA104 skin. These mice promptly rejected the HA104 skin graft with kinetics identical to the directly regrafted Tg⁺ Clone 4 mice (3 of 4 host mice rejected, MST: 9 days, p>0.05 compared with directly regrafted animals). Taken together, these results indicate that rejection of HA104 skin by Clone 4 hosts duplicates a normal physiological immune response with respect to the development of immunological memory.

Having established that Clone 4 hosts consistently reject HA104 skin grafts, we next established an adoptive transfer model in the CD8 transgenic system analogous to that described previously for the CD4⁺ TS1 system [14, 15]. The adoptive transfer model has a number of advantages. Most notable are that rejection is mediated by a more physiological number of T cells and that labeling graft-reactive cells prior to transfer allows evaluation of their activation in vivo. This model represents a refinement over previous investigations as the transgenic transferred T cells are required to mediate rejection as non-transgenic animals do not reject HA⁺ skin (Fig. 1).

As we had previously demonstrated that 5×10⁵ adoptively transferred CD4⁺ T cells from TS1 hosts consistently reject HA104 skin grafts [14], we determined whether this number of Clone 4 lymphocytes would also mediate rejection. 5×10⁵ unfractionated Clone 4 lymphocytes transferred into BALB/c recipients grafted one day prior with HA104 skin consistently mediated rejection with a tempo similar to that observed in directly transplanted Clone 4 Tg⁺ recipients (3/4 rejected, MST: 18.3 days; Fig. 1). Further investigation suggested that 5×10⁵ transferred cells was near optimal for rejection as progressively decreasing the transferred cell number to 7.5×10⁴ led to less consistent graft rejection (Fig. 2). We therefore utilized transfers of 5×10⁵ cells for further experiments. Flow cytometric analysis demonstrated that Thy 1.1⁺ Clone 4 lymphocytes constituted on average 0.35% of peripheral CD8⁺ T cells in BALB/c hosts at two weeks following cell transfer (data not shown). Thus, rejection of HA-expressing grafts is maintained when the precursor frequency is decreased to better recapitulate the prevalence of TCRs reactive against minor determinants.

As mentioned, a principle advantage of the adoptive transfer model is the ability to visualize the function and activation of transferred graft-reactive T cells. Specifically, labeling lymphocytes prior to transfer with the fluorescein-based dye CFSE allows visualization and quantitative assessment of proliferation [16]. We next evaluated the in vivo activation of Clone 4-derived CD8⁺ T cells and compared this with our results using CD4⁺ T cells from TCR transgenic TS1 hosts [15].

To analyze the localization and activation of transferred Clone 4 lymphocytes, 5×10⁵ unfractionated CFSE-labeled Clone 4 lymphocytes were transferred to BALB/c hosts grafted one-day prior with HA104 skin. Axillary and inguinal lymph nodes ipsilateral and contralateral to the HA104 skin graft were harvested 14 days post-cell transfer and the proliferation of transferred cells was evaluated at each site using flow cytometry. We observed an extensive proliferative response among transferred cells that was largely restricted to the draining lymph node (>7 rounds of division; Fig. 3A). While many cells divided 6+ times in the draining lymph node, few cells reached 4+ divisions in contralateral lymph nodes (Fig. 3B). Proliferation was restricted to CD8⁺ T cells, as the cells expressing low levels of CD8, a majority of which would be CD4⁺ T cells, underwent very few rounds of division (Fig. 3A, CD8^lo population), providing evidence that the cells proliferating in response to the graft were Clone 4 lymphocytes bearing the transgenic TCR. The observed proliferative response was highly antigen specific, as the frequency of mitotic events amongst HA-specific T cells in the transfer inoculum (best identified by the expression of the transgene-encoded TCR β chain Vβ8.2) was significantly higher than that among non-HA-specific cells (avg. number of mitoses/10,000 cells: 5556 Vβ8.2⁺ vs. 1055 Vβ8.2⁻, n=3, p=0.02; Figs. 3C and D). The extent of antigen-specific proliferation of Clone 4 cells (>7 divisions) as well as the restriction of proliferation primarily to the draining lymph node were comparable to the results observed in our previously established TS1 model (Figs. 3E and F). While comparison of Figs. 3B and F suggests that there may have been slightly more proliferation of Clone 4 CD8⁺ cells in the contralateral lymph node compared with TS1 CD4⁺ cells, this response was highly variable. Whether the proliferation of CD8 T cells is less tightly restricted than that of CD4 cells would require further investigation.

Collectively, these data indicate that transferred CD8⁺ T cells from Clone 4 hosts are dynamically comparable to transferred CD4⁺ T cells from TS1 hosts [14]. Both populations divided extensively to HA-expressing skin grafts, and proliferation in both instances was highly antigen-specific and largely localized to the draining lymph node.

The successful development of this antigen-specific model of CD8-mediated rejection provided the unique opportunity to directly compare the proliferative responses of individual CD4⁺ and CD8⁺ T cell subsets to a common antigen in vivo. We demonstrated that the proliferation of CD8 and CD4 T cells to HA antigen was comparable at 14 days post-transplant. However, this late time point could represent the peak response, and there may still be differences in the kinetics of activation of CD4 and CD8 T cells post-transplant. We therefore quantitatively examined the kinetics of the proliferative response of each of these T cell subsets to determine whether antigen-specific proliferation of the CD8⁺ T cells mediating graft rejection would parallel that observed in the TS1 CD4⁺ T cell subset.

Immunocompetent syngeneic BALB/c host mice grafted with HA104 skin received unfractionated CFSE-labeled TS1 or Clone 4 lymphocytes one day after skin transplantation. Recipients were euthanized 5 and 10 days after cell transfer, and the proliferative response at the draining ipsilateral axillary and inguinal, contralateral axillary and inguinal, and distant cervical lymph node sites was assessed by flow cytometry.

Five days after cell transfer, CFSE-labeled lymphocytes from both TS1 and Clone 4 mice could be detected in the cervical, ipsilateral, and contralateral lymph nodes of all recipient BALB/c host mice. At this early time point, the CFSE-labeled transgenic T cells of either T cell subset had undergone very little division (avg. number of mitoses per 10⁴ events at ipsilateral lymph nodes: 226 for TS1 (n=5) and 360 for Clone 4 (n=3), p=0.44; Fig. 4). Ten days after cell transfer, both the TS1 and Clone 4 transgenic T cell populations demonstrated increased proliferation restricted to the ipsilateral lymph nodes (avg. mitoses per 10⁴ events: 769 for TS1 (n=5) and 2399 for Clone 4 (n=5); Fig. 4). Division for both populations was significantly increased relative to day 5 (p=0.05 for Clone 4, p<0.0001 for TS1). There was no significant difference between the proliferation of Clone 4 and TS1 lymphocytes at day 10 post-transplant (p=0.09) although there was a trend toward a more rapid CD8 response (Fig. 4). Therefore, these data indicate that the similar pattern of activation for TS1 and Clone 4 lymph node cells that was initially observed at day 14 post-transplant extends to the early post-transplant period.

Our data suggest that both CD4⁺ and CD8⁺ T cells proliferate with similar kinetics in vivo to a common alloantigen. To determine whether these cells also acquire effector function in a similar fashion, we stimulated CFSE-labeled Clone 4 and TS1 lymphocytes in vitro with HA peptide for 72 h and measured intracellular perforin and IFN-γ, respectively. Consistent with prior work linking effector function with proliferation [2, 10], IFN-γ was upregulated after multiple rounds of cellular division within the CD4⁺ T cells. Acquisition of effector function in the CD8⁺ compartment mirrored these findings, as perforin protein levels peaked after approximately four rounds of division (Fig. 5). These data suggest that CD4 and CD8 cells not only proliferate with similar kinetics to the same antigen, but also acquire effector function in a similar manner.