It has been hypothesized that rapidly dividing tumor cells can outpace adoptively transferred antitumor lymphocytes when tumors are large. However, this hypothesis is at odds with clinical observations indicating that bulky tumors can be destroyed by small numbers of adoptively transferred antitumor T cells. We sought to measure the relative growth rates of T cells and tumor cells in a model using transgenic CD8+ T cells specific for the gp10025-33 H-2Db epitope (called pmel-1) to treat large, well-established s.c. B16 melanoma. We tested the effect of the immunization using an altered peptide ligand vaccine alone or in combination with interleukin-2 (IL-2) by analyzing the kinetics of T-cell expansion using direct enumeration. We found that pmel-1 T cells proliferated explosively during a 5-day period following transfer. Calculations from net changes in population suggest that, at the peak of cell division, pmel-1 T cells divide at a rate of 5.3 hours per cell division, which was much faster than B16 tumor cells during optimal growth (24.9 hours per cell division). These results clearly indicate that the notion of a kinetic “race” between the tumor and the lymphocyte is no contest when adoptively transferred cells are stimulated with immunization and IL-2. When appropriately stimulated, tumor-reactive T-cell expansion can far exceed the growth of even an aggressively growing mouse tumor. (Cancer Res 2006; 66(2): 1132-8)
Multiple approaches have been applied to cancer immunotherapy (1–6). Adoptive transfer of autologous lymphocytes with antitumor reactivity is one of the potential immunotherapeutic means currently under investigation for the treatment of cancer (7). The scientific foundation for these studies is composed of data from murine tumor systems that model the treatment effects of clinical adoptive therapy. Many of these models include systems using transgenic lymphocytes targeting highly immunogenic model antigens expressed on transduced tumor cell lines (8–11). Other groups have described similar adoptive transfer models, which use transgenic (8) or tumor challenge-derived lymphocytes (12, 13) that target naturally occurring tumor antigens. Together, these scientific reports have shown that adoptive immunotherapy can be successful in treating tumors in murine systems. However, the treatment of large, vascularized tumors is generally successful only when targeting highly immunogenic “model” antigens. Systems using naturally expressed tumor-associated antigens generally are only effective in the treatment of smaller tumors. One potential explanation for the lower efficacy of these systems may lie in the fact that the bulk of these studies has focused on the adoptive transfer of tumor-reactive cells alone, without further activation by immunization or cytokine support.
Based on these models, some have suggested that the adoptive transfer of tumor-specific T cells fails to successfully treat larger, well-established tumor masses because the tumors contain too many rapidly dividing cells (8). This, in turn, was taken to indicate that the biological reason for the failure of adoptive cell transfer in some settings was due to a kinetic disparity between the growth of the malignancy and the accumulation of antitumor effector cells, the interpretation being that tumor cells simply outgrew any antitumor response (8). However, this reasoning is at odds with results of recent clinical trials, which indicate that relatively small numbers of adoptively transferred T cells (1010 cells) can treat tumor volumes that can be measured in kilograms (>1012 tumor cells; refs. 7, 14, 15). These studies have shown that several of the responder patients exhibited profound lymphocytosis within 8 days after receiving the cells (15). In addition, several animal studies have also shown a rapid proliferation of adoptively transferred T-cell response to viral infection (16–18). However, these studies do not address the kinetics of the adoptively transferred T-cell response in the context of its potential for immunotherapeutic treatment of cancer.
In the current clinical trials, polyclonal tumor-infiltrating lymphocytes (TIL) that are widely heterogenous for avidity, activation, and differentiation status are the primary source of cells for adoptive transfer. However, measurement of TIL expansions is greatly complicated by their heterogeneity. Each clonotype can become activated, undergo expansion, and become inactivated or apoptotic at rates that vary widely from one clonotype to another. Thus, the use of a T-cell receptor (TCR) transgenic mouse derived from a single clonotype, minimizes intraclonal diversity and serves as an optimal model to study the kinetics of a T-cell response during an active, antitumor treatment.
We have successfully modeled human adoptive immunotherapy in mice using a transgenic T cell expressing the TCR specific for the mouse melanoma antigen, gp10025-33 H-2Db epitope, called pmel-1. On the adoptive transfer of in vitro cultured pmel-1 cells followed by recombinant fowlpox virus encoding hgp100 (FP-hgp100) immunization and high dose of interleukin-2 (IL-2) administration into sublethally irradiated recipient mice, regression of large cutaneous murine B16 melanoma was achieved (19). This model targets a naturally expressed tumor antigen and therefore provides us a unique opportunity to study the population kinetics of an adoptively transferred self/tumor-reactive CD8+ T-cell response necessary to mediate regression of large tumors.
In this study, we found a clear association of the initial kinetic response of adoptively transferred, self/tumor-reactive T cells and the regression of established B16 melanoma. We also investigated the influence of immunization and IL-2 administration on the kinetics of the transgenic cell population. Finally, we mathematically determined net rates of division at different stages of the T-cell response and compare these growth rates with that of established B16 tumor to investigate the issue of kinetic disparity between tumor growth and T-cell expansion.
Materials and Methods
Animals and cell culture. We obtained C57BL/6 recipient mice from the National Cancer Institute-Frederick (Frederick, MD). pmel-1 TCR transgenic mice have been described previously (13). pmel-1 mice are on a C57BL/6 background and express a gp10025-33 H-2Db epitope-specific TCR. We crossed pmel-1 mice with syngeneic mice expressing the secondary thy1.1 (CD90.1) congenic marker to aid in the detection of transgenic cell populations. To culture pmel-1 T cells for adoptive transfer, splenocytes from pmel-1 transgenic mice were cultured in RPMI with 10% FCS and 60 IU/mL recombinant human IL-2 (Chiron Corp., Emeryville, CA) and stimulated with 1 μg/mL hgp10025-33 peptide for 6 to 7 days. Cells cultured in this fashion result in a >95% pure CD8+ cell population that expresses both the transgenic TCR and thy1.1 markers as measured by fluorescent cell sorting analysis for CD8+, TCR Vβ13+, and thy1.1+ surface staining of the cells.
Adoptive transfer of pmel-1 T cells and in vivo immunization. Before adoptive transfer of cells, we sublethally irradiated C57BL/6 recipient mice at 500 rads the day of adoptive transfer to deplete endogenous T-cell populations that might interfere with subsequent cell analysis of donor T cells. We then delivered cultured pmel-1 T cells or Bg-1 transgenic T cells [specific for Kb-restricted β-galactosidase (β-gal) protein96-103, used as a control] by tail vein injection at a concentration of 1 × 107 cells in 0.5 mL PBS. Following adoptive transfer of pmel-1 T cells, mice are immunized i.v. with 2 × 107 plaque-forming units of recombinant FP-hgp100 or β-gal virus (Therion Biologics, Cambridge, MA) in 0.5 mL PBS. In addition, recombinant human IL-2 was given i.p. (600,000 IU in 0.5 mL PBS twice daily for a total of six injections). For tumor treatment experiments, 1 × 105 B16.F10 melanoma cells were injected s.c. into C57BL/6 mice 14 days before adoptive transfer.
Harvesting and quantifying transgenic T cells. We harvested spleens from two to five mice given identical treatments at selected time points as described in figure legends. The total numbers of splenocytes were derived from a manual cell count of single-cell suspensions, and average splenocyte populations for single spleens were calculated. Samples of each cell population were then stained for surface Vβ13 TCR chain and the congenic marker thy1.1 using FITC-labeled anti-mouse Vβ13 TCR and phycoerythrin anti-mouse CD90.1 antibodies (BD Biosciences, San Diego, CA), respectively, and analyzed by flow cytometry. We determined the total splenocyte count by multiplying the total splenic pmel-1 T-cell population by the percentage of the splenocytes positively stained for Vβ13 and thy1.1.
Depletion of transferred pmel-1 T cells. B16-bearing C57BL/6-recipient mice were irradiated and given standard adoptive transfer of pmel-1 T cells, immunization, and IL-2 treatment described in previous sections. We depleted CD8+ T cells by injection of 150 μg purified anti-mouse CD8α antibody (clone 53-6.7, BD Biosciences) i.p. at indicated time points. Control antibody, purified rat IgG2a κ (BD Biosciences), was injected into recipient mice at the identical amount and concentration as CD8-depleting antibody. Two days after antibody injection, cell staining analysis of harvested splenocytes from depleted mice exhibited >95% depletion of the transgenic cell population (CD8+, thy1.1+).
Bromodeoxyuridine labeling and proliferation assay. Analysis of proliferation was done using a FITC-conjugated bromodeoxyuridine (BrdUrd) flow kit (BD Biosciences). We injected mice i.p. with 1 mg BrdUrd in 0.1 mL PBS 16 hours before harvesting, fixing, and permeabilizing splenocytes with Cytofix/Cytoperm (BD PharMingen, San Diego, CA). We then treated the cells with DNase I to cleave genomic DNA from fixed cells and intracellular staining with FITC-labeled anti-BrdUrd antibodies was done at 4°C for 30 minutes to detect presence of incorporated BrdUrd. We then stained cell samples for the congenic marker thy1.1 as described above to identify the transgenic cell population using flow cytometry. Histograms were gated on thy1.1+ lymphocyte populations.
Rapid expansion of transferred pmel-1 T cells occurs early during tumor treatment. We employed a model of the treatment of the highly aggressive, poorly immunogenic B16 melanoma, which grew and vascularized for 14 days after implantation. We then adoptively transferred pmel-1 T cells alone or together with a recombinant fowlpox virus expressing an altered peptide ligand of the gp10025-33 epitope and/or IL-2 cytokine (19). Although there was evidence of minor tumor regression in mice receiving pmel-1 cells and FP-hgp100 immunization without additional IL-2, maintenance of B16 regression past 2 weeks following treatment was only observed with the full complement of adoptive transfer of the pmel-1 tumor-reactive cells, immunization with FP-hgp100, and IL-2 cytokine administration (Fig. 1A).
To study the kinetics of the tumor-reactive pmel-1 T-cell population during the tumor treatment, we measured the donor T-cell population over a time course of the first 2 weeks following adoptive transfer and immunization. Levels of transferred pmel-1 T cells were between 11,000 and 180,000 cells in the spleen 1 day following adoptive transfer. We found that the transgenic cell population increased in the spleen in all treatment groups during the next 4 to 7 days (Fig. 1B). Despite the observation that transgenic T cells increased in all treatment groups, immunization with the relevant antigen using FP-hgp100 and IL-2 cytokine administration induced the generation of a much larger population of total pmel-1 T cells. The peak of the pmel-1 T-cell populations occurred 5 days following transfer and was >35-fold greater in number compared with the peak population when immunizing with an irrelevant antigen and giving IL-2. This peak was also >24-fold higher compared with the population peak when immunizing with FP-hgp100 without IL-2 cytokine and >29-fold greater than expansion of adoptive transfer of cells alone (which peaked on 8 days following adoptive transfer). Following this period of cell expansion, the transgenic cell population dramatically contracted over a 24-hour period to ∼30% of the peak population. The pmel-1 splenocyte population then gradually decreased after day 6 to reach a level of 3% of the peak value by day 11 and stabilizes at this level during the second week following adoptive transfer and immunization (data not shown). Kinetics from multiple experiments consistently followed a similar pattern of proliferation and contraction, peaking between days 4 and 5 following transfer and immunization (Fig. 1B).
The early expansion of transferred pmel-1 T cells with proper immunization and IL-2 was not limited to the spleen. We have carefully examined multiple tissues and organs, including peripheral blood, lymph node, lung, liver, brain, kidney, eye, and tumor, after the adoptive cell therapy. Kinetics of pmel-1 T cells in these tissues were found to be similar, which indicated that the proliferation of pmel-1 T cells with or without FP-hgp100 and IL-2 immunization was not tissue specific (20).
Number of pmel-1 T cells during expansion correlates with tumor treatment efficacy. To study the link between the number of pmel-1 T cells and their relationship to the efficacy of tumor treatment, we did a dose titration experiment. When B16 melanoma-bearing mice were given increasing numbers of pmel-1 T cells (2 × 104, 2 × 105, 2 × 106, and 2 × 107), the efficacy of tumor treatment was directly correlated with the increasing number of cells transferred (Fig. 2A). Although we have described previously the specificity of pmel-1 T cells to B16 tumor (20), we transferred culture transgenic T cells specific for a Kb-restricted β-gal protein (called Bg-1) then immunized with hgp100 and IL-2. Even large numbers (2 × 106) of non-gp100-reactive Bg-1 cells did not cause B16 tumor regression (Fig. 2A).
To investigate how long the pmel-1 T cells were required for tumor regression, we did experiments to deplete pmel-1 cells at different time points following adoptive transfer and treatment in irradiated recipients. Because >95% of the transferred cultured pmel-1 cells were CD8+ T cells, we used an established CD8 depleting monoclonal antibody in our experiments. Mice given pmel-1 T cells, immunization, and IL-2 were subsequently depleted of their CD8+ cell populations at days 4, 10, and 16 following adoptive transfer using an anti-mouse CD8 antibody. Depletion of the CD8+ pmel-1 cell population during the expansion phase of the response (4 days after transfer) resulted in greatly abrogated tumor regression (Fig. 2B). However, depletion at later points (10 and 16 days after transfer) resulted in progressively decreased interference of tumor treatment. We found that >95% of the transgenic CD8+ cells were depleted within 2 days of antibody injection (data not shown). These results suggested that expansion of tumor-reactive cells was associated with the extended regression of established B16 melanoma.
Increase in transgenic T-cell population is directly associated with cell proliferation. Because either transgenic cell proliferation or migration to the spleen in response to different treatment regimens could cause an increase in cell numbers, we used BrdUrd to assess cell proliferation at different time points following adoptive transfer. Proliferation of the transgenic population seemed to be the greatest at the earliest time point (days 2-3) when >50% of the transferred cells were positive for BrdUrd incorporation (Fig. 3). This proliferation level gradually decreases during the cell expansion phase of cell growth (3-5 days after transfer). After this rapid expansion, the proliferation of transgenic T cells comes to an abrupt halt (see Fig. 3, d5-6). This abrupt decrease in proliferation of the pmel-1 T cells corresponds to a marked cell contraction that begins after day 5 as shown by pmel-1 population measurements (Fig. 1B). Proliferation was primarily limited to transgenic cell population as analysis of nontransgenic cells showed that <5% of nontransgenic cells incorporated BrdUrd at any given time point (data not shown). Proliferation at days 1 and 2 following adoptive transfer was not measured as the total number of pmel-1 T cells was found to be insufficient for similar analysis.
Removal of either IL-2 support or immunization resulted in much smaller levels of proliferation of the transgenic cell population compared with that of cells treated with full immunization and IL-2 at all the different time points measured (Fig. 3). Despite differences in the number of proliferating cells, all treatment groups exhibited a similar pattern of proliferation for the transgenic cell population that correlated with the pmel-1 population kinetics shown in Fig. 1B. Analysis of pmel-1 proliferation by BrdUrd incorporation in other organ systems indicated that there were similar levels of pmel-1 T-cell proliferation occurring in lymph nodes, peripheral blood, and spleen during the first week following transfer (20). Thus, the observed increases in transgenic T-cell population were directly associated with cell proliferation and not merely due to altered patterns of cellular migration.
Transferred pmel-1 T cells proliferate faster than B16 tumor cells during expansion. To determine whether there is a detrimental kinetic disparity between the growth of the malignancy and the accumulation of antitumor effector cells in our tumor treatment model, we compared the growth rates of B16 melanoma and that of tumor-reactive pmel-1 T cells.
We first calculated the net changes in the cell populations during the T-cell expansion and early contraction (Fig. 1B, summarized in Table 1). Assuming there was no cell death during the period of measurement, the net fold increase in cell population would be equal to 2n, where n was equal to the number of divisions. Alternatively, the number of divisions could be determined as the log2 of the net fold increase in cell numbers. Therefore,
|Treatment group or tumor .||d1-3 .||d3-4 .||d4-5 .||d5-6 .||d6-7 .||d7-8 .||d8-11 .|
|Treatment group or tumor .||d1-3 .||d3-4 .||d4-5 .||d5-6 .||d6-7 .||d7-8 .||d8-11 .|
NOTE: Calculated changes in net population between given time points. Values represent net fold changes of splenic population during the time listed. Positive values represent fold increases, whereas negative values represent fold decreases. Fold increase = (population at final time point) / (population at initial time point). Fold decrease = − (population at initial time point) / (population at final time point).
Net increase <5% of initial population.
B16.F10 value represented optimal net volume change (days 12-14) of tumor growing in naive recipients over two 48-hour period.
The assumption of no cell death for the calculations of the net population change, although convenient for our calculations, was an untested hypothesis. However, attempts at measuring in vivo apoptosis using standard Annexin V staining and terminal deoxynucleotidyl transferase–based assays to detect early-stage or late-stage apoptotic events for adoptively transferred pmel-1 T cells were, in our experience, unsuccessful (data not shown). This was possibly due to very efficient in vivo clearance mechanisms of dying lymphocytes triggered by very early apoptotic events (21). As this calculation did not consider cell death, which would necessitate compensation of cell loss to achieve the observed net population changes, the derived doubling times were most likely conservative estimates and would be even faster if cell death were taken into consideration.
For pmel-1 T cells transferred in combination with immunization and IL-2, cells underwent a net 502.5-fold increase (∼9 divisions) during the 48-hour period from days 1 to 3 following transfer. Therefore, at the height of cell proliferation between 1 and 3 days following adoptive transfer, fully activated pmel-1 T-cell division occurred at a rate of ∼5.3 hours per cell division. We did similar calculations from other groups (Fig. 1B and summarized in Tables 1 and 2). During the period from 1 to 3 days after adoptive cell transfer, pmel-1 T cells expanded ∼30-fold and at a calculated rate of 9.8 hours per cell division in response to immunization without IL-2 support. Measurement of the kinetics of transgenic cells in response to IL-2 cytokine without immunization resulted in a more modest 9.6-fold increase in population with a doubling time of 14.7 hours. Normal homeostatic proliferation of the transgenic cell population in irradiated recipients resulted in a slight 2.6-fold increase in cell population during this 48-hour period, resulting in a doubling time of 34.8 hours.
|Treatment group or tumor .||d1-3 .||d3-5 .|
|Treatment group or tumor .||d1-3 .||d3-5 .|
NOTE: Calculated doubling time of transferred pmel-1 T cells or B16 tumor cells during the expansion phase. Values represent average rate of expansion as expressed by hours per single division. Doubling time = 48 hours / log2 (fold population or volume increase).
Net decrease in population. Calculation of doubling time was not applicable.
B16.F10 value represented doubling time (hours per single division) of B16 melanoma cells growing in naive mice over a 48-hour period of optimal growth (days 12-14).
We then calculated proliferation rates for a s.c. established B16 tumor by measuring growth of untreated tumor from multiple experiments. We determined the volume of tumors as that of a half-sphere = (4/3πr3) / 2, where the radius of the sphere was the half of the average of perpendicular diameters. Because the average change in tumor volume was approximately proportional to the number of tumor cells comprising the growth, we calculated that the greatest change in tumor growth for naive mice was ∼3.8-fold over a 48-hour period (between days 12 and 14 following tumor challenge; Fig. 4A). B16 tumor growth in naive mice was thus at an approximate rate of 24.9 hours per cell division during optimal growth (Table 2).
We sought to compare the relative rates of T cells and tumor cells and plotted the growth of T cells or tumor on a logarithmic scale as a linear function of time (Fig. 4). We derived the rate at which the cells grew by using the slopes of population changes during different periods. Comparisons of the maximum slopes (at 1-3 days for pmel-1 T cells and 12-14 days for B16 tumor) provide clear evidence that the rate of change of the pmel-1 population in combination with immunization and IL-2 was much greater than that of B16 tumor growth. Additional analysis from doubling time calculations also provided further evidence that the expansion of pmel-1 T cells with full treatment was ∼4.7-fold faster than that of growing B16 tumor and resulted in an ∼130-fold larger increase in total cell population compared with B16 growth over a 48-hour period of time (Tables 1 and 2). Thus, the expansion rate of cytokine and antigen-stimulated adoptively transferred CD8+ T cells exceeds that of the aggressively growing mouse tumor, B16.
In this study, we report on the kinetic response of adoptively transferred transgenic CD8+ T cells following immunization and IL-2 cytokine administration in the context of the treatment of an established s.c. B16 melanoma in recipient mice. The full activation of transferred cells by immunization with FP-hgp100 and IL-2 cytokine resulted in a >30-fold larger peak population than when transferring cells alone, a >35-fold greater population compared with immunizing with an irrelevant antigen and IL-2 support, and a >24-fold higher population peak compared with immunization alone without IL-2 cytokine support. Depletion of the cells at the peak of expansion significantly diminished the efficacy of B16 tumor treatment. In addition, variation in the number of pmel-1 cells used with vaccination and IL-2 directly correlated with antitumor efficacy.
From both population kinetics and proliferation analysis using BrdUrd incorporation, it is clear that most of the cell expansion occurs during the first three days following adoptive transfer. However, net changes in the pmel-1 population during this period must also take in consideration cell redistribution and migration to the spleen. Immunization and IL-2 could alter the migration and survival of adoptively transferred pmel-1 T cells, especially in the period just after adoptive transfer. Studies on the early kinetics of endogenous lymphocytes in response to superantigen SEB have shown that CD8+ cell populations were transiently decreased in the blood and spleen and were increased in the lymph nodes during antigen presentation and cell activation within the first 24 hours (22). Following the first 24 hours, there was rapid equilibration of the cells among the three compartments. In our own studies, we started measurements of T-cell distribution after the first 24 hours; we did not see evidence of selective distribution of the pmel-1 cells (23).
When describing the T-cell kinetic response, we must draw comparisons with previously described model systems using similar TCR transgenic T cells to follow their progress in response to viral or nonself antigenic stimulation. Using transgenic P14 T cells that were specific for a Db LCMV epitope, several studies from Blattman et al. have defined both endogenous (24) and adoptive transfer (16) responses to LCMV infection. The endogenous response measured in the Blattman et al. study showed an explosive expansion of antigen-reactive T cells following LCMV infection resulting in a calculated >50,000-fold net increase in the reactive cell population during the 8-day expansion phase. Their study further concluded that the endogenous and donor T-cell responses were similar as evidenced by results that naive donor P14 cells could similarly expand >1,000-fold (16). A separate study using transgenic T cells reactive to the male antigen (17) also showed a 51-fold net increase of the transferred cell population following immunization. More recently, van Stipdonk et al. revealed that adoptively transferred ova-specific T cells underwent rapid proliferation on brief antigen stimulation (25). The expansion observed in our B16 tumor treatment model also reflected a very rapid expansion in response to immunization. Despite the differences in total numbers and expansion potential, these models all clearly show that the T-cell response to viral infection and immunization occurs at an extremely rapid pace.
Our studies with the addition of IL-2 administration in combination with immunization showed significant improvement on overall T-cell kinetics and tumor treatment efficacy compared with cells alone. Blattman et al. have studied the endogenous T-cell response to LCMV infection in combination with IL-2 administration. The conclusion of these studies was that administration of IL-2 during the expansion phase of an antiviral immune response was detrimental to the CD8+ T-cell response resulting in an increase in cell death during the contraction phase and had no effect on the overall cell expansion (24). Our own results, using a 80 times higher dose of IL-2 (600,000 IU twice daily), suggested that exposure to the cytokine following immunization resulted in a much greater level of proliferation resulting in a population peak that is >24-fold higher than without cytokine. In terms of cell depletion, our study also showed an increase in the rate of cell death during the contraction phase with IL-2 administration (Table 1). However, the final transgenic cell count at day 11 was found to be greater in recipients receiving IL-2 presumably as a result of the differences in the total peak population at day 5 (Fig. 1B). Previous studies have also suggested that IL-2 is quickly excreted from the body within 2 hours following i.p. injection in mice (26). This rapid cytokine depletion may render an exogenously given IL-2 at the dose of 15,000 IU per injection ineffectual to influence the overall T-cell proliferative response. Our own investigation into dosing of IL-2 has suggested that even at 150,000 IU (one fourth of our standard dose) the efficacy of tumor treatment is unaffected (data not shown).
Although many investigators have reported previously on the efficacy of adoptive transfer models to treat tumors in mice, most of these studies have only addressed the efficacy of the transferred T cells (8, 11, 12, 27–29). Arguments for the use of adoptive cell transfer alone have suggested that it alleviates potential concerns of detrimental and complicating use of adjuvant, cytokines, or immunization. However, in the final analysis, the most compelling argument for using other factors to complement cell transfer is its clear ability to increase the efficacy of tumor treatment. Results from this study clearly showed that complementation of adoptive cell therapy with immunization and IL-2 resulted in a profound difference in treatment of well-established B16 melanoma, which was mediated by the generation of a greatly increased tumor-reactive cell population.
Overcoming the continuous growth of an established tumor mass is the primary goal for adoptive immunotherapy. Speculation by Hanson et al. has suggested that for treatment of large tumors there “may be an insufficient tumor-specific T-cell response, which results in a detrimental kinetic disparity between the growth of the malignancy and the accumulation of host antitumor effector cells” (8). Considering the results from our study, the rate of expansion for adoptive transfer of cells alone was approximated to be slightly slower than the growth of untreated B16 tumor. Thus, it is not surprising that only small tumors could be successfully treated in studies investigating cell transfer alone (8). The rate of T-cell expansion after adoptive cell transfer with immunization and IL-2 far exceeded the rate at that even the highly aggressive B16 tumor could grow. Our analysis of growth rates indicated that the expansion differential of pmel-1 T-cell proliferation with full treatment was occurring ∼4.7-fold faster than that of growing B16 tumor and resulted in an ∼130-fold larger increase in total cell population compared with B16 growth over a 48-hour period. Although this initial rate of expansion by T cells was sufficient to destroy tumors in these settings, high levels of antitumor T cells were not sustained after the first week. We therefore believe that one of the keys to the improvement of adoptive therapy for cancer will depend on finding methods to sustain or prolong the proliferation and activity of the tumor-reactive cells.
We have observed a >50% response rate in patients with metastatic melanoma who received cultured tumor-reactive lymphocytes along with high dose of IL-2 in the lymphodepleted setting (14, 15). In this study, we have shown clearly that the successful use of adoptively transferred T cells requires proper in vivo activation of the transferred T cells (19). However, in clinical trials where bulk TILs are transferred, it would be difficult to measure how a heterogeneous population with varieties of antigen specificities would respond to effective stimulation. Interestingly, in some of the responders, lymphocytosis was observed and presumably induced by endogenous tumor antigen.
In summary, the observations in this report clearly indicate that adoptive cell transfer accompanied by immunization and IL-2 can result in a dramatic increase in proliferation of the transferred T-cell population. This increased population of antitumor T cells can mediate the extended regression of well-established B16 melanoma. Our results also suggest a need for additional measures to sustain the high level of tumor-reactive T cells following the initial proliferation response. One possibility to improve the sustenance of transferred T cells is the substitution of IL-2 with other growth cytokines, such as IL-15, which may not induce increased cell death or the induction of T regulatory cells that is attributed to IL-2 during cell contraction (30). Alternatively, continuous immunization against the targeted tumor-associated antigens and cytokine exposure may serve to reactivate the contracting T-cell population. Analysis of T-cell kinetics is but one aspect correlating to the efficacy of the adoptive cell therapy and provides us with a model by which we may be able to rationally design and assess future strategies in clinical adoptive cell therapies.
Note: L.N. Hwang and Z. Yu contributed equally to this article.
Grant support: Intramural Research Program of the NIH, National Cancer Institute and in partial fulfillment of a Ph.D. in Biochemistry (D.C. Palmer) at the George Washington University (Washington, DC).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.