Depletion of tumor-induced regulatory T cells prior to reconstitution rescues enhanced priming of tumor-specific, therapeutic effector T cells in lymphopenic hosts

Abstract

We reported previously that vaccination of reconstituted-lymphopenic mice resulted in a higher frequency of tumor-specific effector T cells with therapeutic activity than vaccination of normal mice. Here we show that lymphopenic mice reconstituted with spleen cells from tumor-bearing mouse (TBM), a situation which resembles the clinical condition, failed to generate tumor-specific T cells with therapeutic efficacy. However, depletion of CD25⁺ Treg from the spleen cells of TBM restored tumor-specific priming and therapeutic efficacy. Adding back TBM CD25⁺ Treg to CD25^- naïve and TBM donor T cells prior to reconstitution confirmed their suppressive role. CD25⁺ Treg from TBM prevented priming of tumor-specific T cells since subsequent depletion of CD4⁺ T cells did not restore therapeutic efficacy. This effect may not be antigen-specific as three histologically distinct tumors generated CD25⁺ Treg that could suppress the T cell immune response to a melanoma vaccine. Importantly, since ex vivo depletion of CD25⁺ Treg from TBM spleen cells prior to reconstitution and vaccination fully restored the generation of therapeutic effector T cells, even in animals with established tumor burden, we have initiated a translational clinical trial of this strategy in patients with metastatic melanoma.

Introduction

Adoptive immunotherapy with tumor-specific effector T cells (T_E) has been shown to be a highly effective therapeutic modality in animal tumor models. One approach to generate tumor-specific T_E for adoptive transfer is to activate and expand tumor-vaccine draining lymph node (TVDLN) T cells. T_E generated from TVDLN exhibit tumor-specific function in vitro and therapeutic efficacy in vivo [1]. CD8⁺ T cells are the primary effector cells; however, CD4⁺ T cells are required to maintain long-term memory and cure treated hosts [1, 2]. The translation of this strategy to the clinic has generated some objective clinical responses [3, 4], but overall results have been disappointing [4-7].

The development of effective adoptive immunotherapy is a complex process, during which there are multiple steps, the failure to accomplish any of which may explain the failure of this treatment strategy. The major steps include the generation of tumor-specific CD8⁺ and CD4⁺ T cells before adoptive transfer, and the expansion, trafficking and survival of tumor-specific, therapeutic T cells following adoptive transfer.

However, the initial barrier limiting efficacy is the generation of sufficient tumor-specific T cells with therapeutic potential in a host with progressive tumor burden. Recent advances in our understanding of lymphocyte homeostasis and the effects of the lymphopenic environment on T cell responsiveness to antigen stimulation have provided a rationale for combining lymphodepletion with vaccination. Mackall and colleagues first documented that antigen administration during homeostasis-driven expansion led to a dramatic skewing of the T cell repertoire towards T cells specific for the administered antigen [8]. Subsequent studies showed that cancer vaccines administered during recovery from a lymphopenic state resulted in augmented immune responses to tumor-associated antigens and improved therapeutic efficacy.

The repeated administration of tumor lysate-pulsed DC vaccines shortly after bone marrow transplantation [9], tumor vaccines administered in mice with mixed allogeneic bone marrow chimeras [10], and the transfusion of autologous or syngeneic T cells into sub-lethally irradiated lymphopenic mice [11], led to growth inhibition of previously established tumors cells.

In our studies, vaccination during homeostasis-driven proliferation increased the number of tumor-specific CD8⁺ and CD4⁺ T cells in reconstituted, lymphopenic mice (RLM), which exhibited increased tumor-specific type 1 cytokine secretion and therapeutic efficacy in both adoptive immunotherapy [12] and active-specific therapy studies [13].

Previously, we showed that in addition to improving the direct anti-tumor effects, tumor-specific CD4⁺ T cells appeared to be critical for the survival and curative capacity of tumor-specific CD8⁺ effector T cells. Interestingly, the adoptive transfer of tumor-specific CD8+ effector T cells in combination with exogenous IL-2 was sufficient to mediate tumor regression in normal animals transiently depleted of CD4+ T cells by mAb treatment; however, when these studies were repeated in MHC II KO mice that lacked classical class II-restricted CD4⁺ T cells, pulmonary metastases could be eliminated, but all animals subsequently died of new metastatic disease. This supported an important role for CD4⁺ T cells in the surveillance and/or control of metastatic disease after the adoptive transfer of tumor-specific T_E [2].

There is a growing number of studies showing that the proliferation and efficacy of tumor-reactive CD8⁺ T cells can be inhibited by CD4⁺CD25⁺ Treg [14, 15, 16]. Turk et.al. showed that concomitant tumor immunity to a poorly immunogenic melanoma was prevented by Treg [17], with the underlying mechanism being CD4⁺ T helper cell inhibition and prevention of CD8⁺ T cell maturation [18]. Furthermore, immune responses to peptide vaccines in patients with cancer were enhanced when Treg were depleted [19] and preclinical studies combining CD25 depletion and homeostatic proliferation improved the generation of a potent antitumor immune response [20, 21]. These studies suggest that tumor-induced CD4⁺CD25⁺ Treg could explain the failure of vaccines to generate therapeutic T_E in tumor-bearing hosts [22].

Herein we examine the influence of tumor-induced regulatory CD4⁺ T cells on the priming of effector T cells following reconstitution of a lymphopenic, tumor-vaccinated host and show that depletion of CD25⁺, but not CD4⁺ T cells, from donor TBM spleen cells in our clinically applicable RLM model restored the generation of therapeutic T cells. Importantly, vaccination with tumor cells was also effective when the vaccinated reconstituted, lymphopenic host was carrying a large, systemic tumor burden.

Results

Reconstitution and vaccination of an irradiated, lymphopenic host primes a higher frequency of tumor-specific T cells

We previously reported that vaccination of congenitally lymphopenic, Rag-1^-/- mice reconstituted by infusion of naïve spleen cells resulted in an augmented anti-tumor response in the TVDLN. Following in vitro activation and expansion, TVDLN contained higher frequencies of tumor-specific CD8⁺ and CD4⁺ T cells, which were significantly more therapeutic in adoptive immunotherapy studies than T_E generated from intact naïve mice [12]. Here we asked whether reconstitution and vaccination of mice made lymphopenic by irradiation with 500 rads (500R) would also result in the generation of an increased frequency of tumor-specific T cells in TVDLN. Effector T cells generated from TVDLN of irradiated mice that were reconstituted with 20 × 10⁶ spleen cells and vaccinated with D5-G6 contained a significantly increased frequency (p<0.05) of IFN-γ-producing (Figure 1A) and TNF-α-producing (Figure 1B) tumor-specific T cells, respectively. These T_E were also substantially more cytolytic than T_E generated from “intact”, non-RLM (Figure 1C). These results demonstrate improved priming of tumor-specific T_E in TVDLN from reconstituted hosts, which received radiation to induce lymphopenic conditioning. Parallel experiments also used chemotherapy (cyclophosphamide) to induce lymphopenia, which resulted in equal efficacy to induce lymphopenia in this model (data not shown).

Figure 1. Cytokine secretion and cytotoxicity of T_E from reconstituted lymphopenic mice.

T_E were isolated from TVDLN of wt and wt500R mice vaccinated with D5-G6 for 8 days and stimulated for 12h in CM alone (None), with tumour cell lines D5, MCA-310, D5^CIITA, and MCA-310^CIITA, or with immobilized anti-CD3 mAb (data not shown). The percentage of IFN-γ⁺ (A) or TNF-α⁺ (B) T_E of the total CD8⁺/CD4⁺ CD3⁺ T_E (gated 50 000 events) is presented. Data show mean ± SEM from three (IFN-γ) and two (TNF-α) independent experiments and significant increases (*p<0.05; two-tailed t-test) of cytokine expression between wt (black square) and wt500R (white square) T_E for each stimulation are indicated. (C) T_E from wt (triangles) or wt500R (circles) mice were stimulated either with 1 × 10^{4 51}Cr-labeled D5 (black) or syngeneic MCA-310 tumor cells (white) as controls for 6h at the indicated effector to target ratios (E/T). Data shown is representative of two independent experiments.

Effector T cells generated in lymphopenic hosts reconstituted with spleen cells from tumor-bearing mice are non-therapeutic

Our preclinical studies suggest that vaccination of reconstituted lymphopenic patients would augment their response to cancer vaccines; however, a major difference between the preclinical studies and their clinical application is that the lymphocytes for reconstitution would have to come from a cancer-bearing patient rather than a naïve “non-tumor-bearing” donor. Here we sought to examine whether reconstitution of animals with spleen cells from mice with metastatic disease would still generate therapeutic T_E. Mice were injected i.v. with tumor cells and 8 to 11 days later, their spleen cells were used for reconstitution. T_E generated from mice reconstituted with spleen cells from tumor-bearing mice (TBM) exhibited reduced or no therapeutic efficacy in all experiments (Table 1) and failed to secrete tumor-specific IFN-γ in intracellular cytokine and release assays (data not shown).

Table 1. Advanced systemic tumor burden inhibits priming of tumor-specific T_E.

			Mean No. D5 pulm. metastases (SEM)
RLM Donor	TE	IL -2	Exp 1	Exp 2	Exp 3	Exp 4
None	None	+	> 250	> 250	> 250	> 250
Naïve	+	+	0*	0*	24 (11)*	32 (14)*
8d TBM	+	+	99 (31)*†	nd	nd	nd
11d TBM	+	+	nd	nd	> 250	> 250
14d TBM	+	+	nd	> 250	> 250	> 250

Lymphopenic mice were reconstituted with spleen cells donated by naïve, 8, 11, or 14-day TBM, vaccinated and 20 ×10⁶ effector T cells generated from TVDLN used in adoptive immunotherapy of animals bearing 3-day D5 pulmonary metastases. On day 14, lungs were harvested, fixed in Fekete's solution and pulmonary metastases were counted. Data are from 4 consecutive, independent experiments. Significant tumor regression (*p<0.05) in comparison to controls and significantly reduced (†p<0.05) therapeutic efficacy in comparison for 8-day TBM vs. naïve T_E was analyzed by comparing the number of pulmonary metastases by Wilcoxon rank sum test.

Reconstitution with CD25-depleted spleen cells restores priming of tumor-specific T cells

CD4⁺ Treg have been identified in mice and patients with cancer [14, 15, 16, 23-25]. We speculated their presence may explain the failure to generate therapeutic T_E from mice reconstituted with spleen cells from TBM. We posited that elimination of these cells prior to reconstitution would eliminate the suppression and promote the generation of therapeutic T_E. Previously, we reported that vaccination of naïve, intact animals with a GM-CSF secreting tumor vaccine, D5-G6, can efficiently prime therapeutic, tumor-specific CD8⁺ T_E independent of CD4⁺ T cell help [2]. On that basis, we speculated that reconstitution with CD4-depleted TBM spleen cells would avoid the suppressive effects of CD4⁺ T cells and increase the priming of tumor-specific CD8⁺ T cells. Spleen cells from naïve as well as 11 and 14 day TBM mice were used for these studies. CD4⁺ cells were depleted by either in vivo administration of anti-CD4 mAb prior to spleen cell harvest or in vitro by treatment of spleen cells with anti-CD4 magnetic beads and passage over a magnetic column. Following treatment, control and CD4-deleted spleen cells were immediately used to reconstitute 500R-irradiated mice. Not only did the depletion of CD4⁺ T cells fail to improve therapeutic efficacy (Groups 5-10, Supplemental table 1), but depletion of CD4+ T cells actually reduced or prevented priming of therapeutic T cells (Supplemental Table 1). Thus, these findings indicate, that lymphopenic conditioning of the host changes the role of CD4⁺ T cell help in this model.

While initial experiments indicated a higher frequency of CD4⁺ CD25⁺ T cells in TBM compared to naïve spleens, a series of more than 10 independent analyses found no significant difference in the frequency or phenotype of CD4⁺CD25⁺ T cells in naïve and 8-14 day TBM. Nonetheless, we posited that there may have been a functional change in the TBM CD25⁺ cells that suppressed T cell priming in RLM. Therefore, CD25⁺ T cells were magnetically eliminated (purity >98.5%, Figure 2A) from the spleen cells used to reconstitute lymphopenic animals (Experimental Design, Figure 2B). To further address the role of CD25⁺ cells from TBM, 10⁶ cells of the purified CD25⁺ population were added back to the CD25-depleted naïve or TBM spleen cells used to reconstitute lymphopenic mice (Figure 2B). All mice were then vaccinated with D5-G6 and 8 days later TVDLN were harvested to generate T_E. All groups were treated in parallel.

Figure 2. In vitro CD25-depletion from naïve and TBM spleen cells and “add back” of CD25⁺ spleen cells from TBM before reconstitution.

(A) Schematic protocol of CD25-depletion of spleen cells used for reconstitution of lymphopenic hosts. 20 × 10⁶ fresh naïve or 8-14 days systemic D5 (pulmonary metatstases) tumor-bearing mouse (TBM) spleen cells were used to reconstitute lymphopenic hosts, irradiated with 500 R (wt500R). CD25⁺ cells were magnetically depleted (MACS). Total or CD25-depleted spleen cells were used to reconstitute lymphopenic mice. In indicated experiments, 1 × 10⁶ CD25-enriched TBM spleen cells were “added back” to the CD25-depleted naïve and TBM spleen cell populations prior to reconstitution (total 20 × 10⁶ spleen cells). Reconstituted, lymphopenic mice (RLM) were immediately vaccinated with 10⁶ D5-G6 tumor cells injected subcutaneously in all four flanks. In some experiments, non-RLM wt mice were vaccinated as controls. Eight days following tumor inoculation the tumor vaccine draining lymph nodes (TVDLN) were harvested, single cell suspensions generated, washed and activated for two days in the presence of anti-CD3 and anti-CD28 mAb. The T cells were harvested and expanded for three days in the presence of low-dose IL-2. The resulting effector T cells were adoptively transferred into wt mice bearing 3-day established pulmonary metastases and used parallel in in vitro assays. In some experiments, mice bearing 3-day D5 pulmonary metastases were irradiated (500R) just prior to adoptive effector T cell transfer. Mice were sacrificed 14 days following tumor inoculation, lungs were resected, fixed in Fekete's solution and the number of pulmonary metastases enumerated.

(B) Spleens from naïve or 8-day systemic D5 tumor-bearing mice were harvested and magnetically sorted (MACS). CD25-depleted, CD25-enriched and “add back” spleen cell samples from the populations transferred for reconstitution were evaluated via flow cytometry after staining with anti-CD3 and anti-CD4. Black numbers indicate the % of CD4⁺CD25⁺ and CD4⁺CD25^- T cells from all cells measured, red numbers represent the % of CD4⁺CD25⁺ T cells over all CD4⁺ T cells (black gate boxes for “CD25^pos” and “CD25^pos plus TBM CD25^pos” eliminate background). Dot plots contain 10⁴ CD3⁺ events and are representative of 4 experiments.

In four consecutive experiments, the adoptive transfer of T_E generated from lymphopenic recipients that received CD25-depleted TBM spleen cells showed significant (p<0.05) recovery of therapeutic efficacy. Reconstitution with CD25-depleted naïve spleen cells also resulted in therapeutic T_E activity (Table 2). Adding back TBM CD25⁺ spleen cells to CD25-depleted naïve or TBM spleen cells at the time of reconstitution significantly (p<0.05) reduced therapeutic efficacy of T_E generated from both groups (Table 2). These data provide strong evidence that D5 tumor-induced CD4⁺CD25⁺ T cells from TBM are potent suppressors of the priming of tumor-specific T_E and that their elimination prior to reconstitution recovers.

Table 2. Reconstitution of lymphopenic mice with CD25-depleted TBM spleen cells eliminated tumor-induced suppression and restored therapeutic efficacy of T_E.

				Mean No. D5 pulm. metastases (SEM)
RLM Donor	Host	TE	IL-2	Exp 1	Exp 2	Exp 3	Exp 4
None	wt	None	+	>250	> 250	> 250	> 250
Naïve total	wt	+	+	117 (98)*	5 (4)*	0*	8(8)*
Naïve CD25depl	wt	+	+	193 (40)*	7 (7)*	0*	42(19)*
TBM total	wt	+	+	>250	234 (25)	188(33)*	233(22)
TBM CD25depl	wt	+	+	76 (67)*#	50 (27)*#	0*#	22(29)*#
Naïve CD25depl +TBMCD25pos	wt	+	+	nd	nd	43(27)*†	196(43)†
TBM CD25depl +TBMCD25pos	wt	+	+	nd	nd	116(29)*†	228(24)†

Effector T cells were generated from 8-day tumor-vaccine-draining lymph node cells after in vitro activation/expansion. TVDLN were harvested from wt500R lymphopenic mice, reconstituted with 20×10⁶ total or in vitro CD25⁺-depleted spleen cells from naïve or 8-day TBM mice and immediately vaccinated with D5-G6. Each group represents 5 wt normal animals with established 3-day D5 pulmonary metastases, controls received only 90,000 IU IL-2 i.p. for 4 consecutive days, all other groups shown received 20×10⁶ T_E (Exp.1&2) or 30×10⁶ (Exp.3&4) from indicated RLM donors including IL-2 i.p. On day 14, lungs were harvested, fixed in Fekete's solution and pulmonary metastases were counted. Data represents 4 consecutive experiments. Significant tumor regression (*p<0.05) comparing treatment groups to the non-adoptive transfer controls (None), (#p<0.05) comparing TBM CD25^depl vs. TBM total and (†p<0.05) comparing TBM CD25^depl vs. Naïve CD25^depl+TBM CD25^pos or TBM CD25^depl + TBM CD25^pos was analyzed by comparing the number of pulmonary metastases by Wilcoxon rank sum test.

We also examined the effect of CD25⁺ T cells on tumor-specific cytokine secretion. T_E generated from mice reconstituted with total or CD25-depleted naïve spleen cells demonstrated tumor-specific IFN-γ production; however, the addition of CD25⁺ TBM spleen cells to the CD25-depleted naïve spleen cells eliminated tumor-specific IFN-γ production (Figure 3A). As expected, T_E generated from mice reconstituted with total TBM spleen cells did not produce tumor-specific IFN-γ following stimulation, but depletion of CD25⁺ cells from the population used to reconstitute lymphopenic mice significantly (†p<0.05) restored IFN-γ production in T_E from TVDLN. Importantly, this recovery was blocked almost completely (*p<0.05) when CD4⁺CD25⁺ TBM T cells were added back to the CD25-depleted TBM spleen cells used for reconstitution (Figure 3B). It is noteworthy that the number of CD25-enriched TBM spleen cells added back to naïve CD25-depleted spleen cells was approximately 1/3 of the number of CD4⁺CD25⁺ cells present prior to depletion in the TBM total spleen cell population; 38% and 31% for experiments three and four respectively (Figure 3C; Table 2). The suppressive effect of TBM CD25+ Treg is further illustrated by the photographs of D5 pulmonary metastases remaining 13 days after adoptive transfer of T_E in the described experiments (Figure 3D).

Figure 3. Inhibition of priming a therapeutic anti-tumour response by CD25⁺ cells from TBM.

(A) CD25⁺ spleen cells from TBM inhibit T_E IFN-γ expression in mice reconstituted with naïve spleen cells. Tumor-specific T_E were generated by vaccination with D5-G6 in lymphopenic mice; lymphopenic mice were reconstituted with either total or CD25-depleted naïve spleen cells or CD25-depleted naïve cells mixed with CD25⁺ spleen cells from TBM. Two million T_E were harvest after in vitro activation/expansion and stimulated with D5 tumor cells. As controls, T_E were stimulated with MCA-310 tumor cells, and without any stimulation in CM only. The numbers in the upper right corner of each dot plot show the % of tumor-specific IFN-γ⁺ cells from CD8⁺CD3⁺ T_E, and are representative of 2 independent experiments. (B) CD25⁺ spleen cells from TBM inhibit T_E IFN-γ expression in mice reconstituted with TBM spleen cells. Tumor-specific T_E were generated in lymphopenic mice by vaccination with D5-G6; lymphopenic mice were reconstituted with either total or CD25-depleted TBM spleen cells or CD25-depleted TBM mixed with TBM CD25⁺ spleen cells. T_E were stimulated as described in (A). Data are representative of two independent experiments.

(C) CD25⁺ spleen cells from TBM inhibit the secretion of tumor-specific IFN-γ. T_E shown in Figure 3 A and B were cultured in the absence of Brefeldin A. Supernatants were collected and IFN-γ (pg/ml) secretion was determined by ELISA. Data show mean ± SEM of duplicates from 2 independent, pooled experiments (data in log-scale). Statistically significant decrease in IFN-γ secretion between naïve total, D5-stimulated effector T cells and all other groups (*p<0.05) and the significant increase in IFN-γ secretion between TBM total and TBM CD25depl (†p<0.05) was analyzed using a two-tailed t-test.

(D) Therapeutic efficacy of RLM T_E with or without CD25-depletion in normal tumor-bearing mice. Lungs of wt mice with established 3-day D5 pulmonary metastases, 14 days after tumor inoculation. As control, animals (n=5) received 90,000 IU IL-2 i.p. for 4 consecutive days without T_E transfer. In all other groups (n=5), adoptive transfer of T_E generated in lymphopenic mice, reconstituted with naïve or TBM spleen cells with or without CD25-depletion or “add back” of TBM CD25⁺ T cells was administered including IL-2 i.p. when 3-day pulmonary metastases of D5 were established. On day 14 lungs were harvested, fixed in Fekete's solution and pulmonary metastases are numbered. Data is representative of 4 independent experiments.

Next, we investigated whether induction of lymphopenia prior to the adoptive transfer would increase therapeutic efficacy. In two of four consecutive experiments T_E generated from CD25-depleted TBM RLM showed significantly (#p<0.05) enhanced therapeutic efficacy when transferred into 500R irradiated tumor-bearing recipients as compared to the same T_E transferred into non-irradiated tumor-bearing recipients (Table 2 vs. 3).

Table 3. Therapeutic efficacy of T_E generated from CD25-depleted TBM spleen cells can be restored and maintained in reconstituted, lymphopenic tumor-bearing hosts.

				Mean No. D5 pulm. metastases (SEM)
RLM Donor	Host	TE	IL-2	Exp 1	Exp 2	Exp 3	Exp 4
None	wt500R	None	+	> 250	> 250	239 (18)	> 250
Naïve total	wt500R	+	+	36 (31)*♣	1 (1)*♣	0*♣	0*♣
Naïve CD25depl	wt500R	+	+	1 (1)*♣	0*♣	0*♣	0*♣
TBM total	wt500R	+	+	133 (45)*♣	206 (52)	169 (51)*	238 (15)
TBM CD25depl	wt500R	+	+	0*#♣	0*#♣	2 (1)*#	45 (36)*#
Naïve CD25depl +TBMCD25pos	wt500R	+	+	nd	nd	2 (2)*	64 (28)*†
TBM CD25depl +TBMCD25pos	wt500R	+	+	nd	nd	4 (4)*	178 (28)†

T_E were generated from 8-day tumor-vaccine-draining lymph node cells after in vitro activation/expansion. TVDLN were harvested from wt500R lymphopenic mice, reconstituted with 20×10⁶ total or in vitro CD25-depleted spleen cells from naïve or 8-day TBM mice and immediately vaccinated with D5-G6. Each group represents 5 wt animals with established 3-day D5 pulmonary metastases that received 500R irradiation (wt500R) to induce lymphopenia on the day of adoptive T_E transfer. Controls (n=5) received only 90,000 IU IL-2 i.p. for 4 consecutive days, all other groups (n=5) shown received 20×10⁶ T_E (Exp.1&2) or 30×10⁶ (Exp.3&4) from indicated RLM donors including IL-2 i.p. On day 14, lungs were harvested, fixed in Fekete's solution and pulmonary metastases were counted. Data represents 4 consecutive experiments. Significant tumor regression (*p<0.05) comparing treatment groups to the non-adoptive transfer controls (None), (#p<0.05) comparing TBM CD25^depl vs. TBM total and (†p<0.05) comparing TBM CD25^depl vs. Naïve CD25^depl+TBM CD25^pos or TBM CD25^depl + TBM CD25^pos was analyzed by comparing the number of pulmonary metastases by Wilcoxon rank sum test. Additionally overall-therapeutic efficacy has improved significantly (♣p<0.05) when T_E are adoptively transferred into lymphopenic, (wt500R) tumor-bearing mice as compared to the corresponding groups of T_E transferred into normal tumor-bearing wt mice. Statistical analysis of the number of pulmonary metastases was performed by Wilcoxon rank sum test.

Reconstitution with CD25-depleted spleen cells is also effective in lymphopenic hosts with systemic tumor burden

For a CD25-depletion strategy to be effective in cancer patients, vaccination during lymphopenia must be effective in a host with metastatic disease, not just a host that has received CD25-depleted lymphocytes from a tumor-bearing donor. To model this situation, we attempted to prime CD25-depleted TBM spleen cells by vaccination of lymphopenic hosts bearing systemic tumor burden. As a positive control, naïve or tumor-bearing lymphopenic mice were reconstituted with naïve spleen cells. When naïve spleen cells were used to reconstitute lymphopenic mice, the vaccination strategy primed tumor-specific, therapeutic T cells regardless of whether the recipient was naïve or had systemic tumor metastases (Figure 4A and 4B). Consistent with data presented above, T_E generated from RLM receiving total TBM spleen cells failed to exhibit tumor-specific release of IFN-γ (Figure 4A) and lacked therapeutic function in adoptive immunotherapy experiments (Figure 4B), regardless of whether the vaccinated, lymphopenic mouse was tumor free (naïve) or had tumor metastases. Importantly, reconstitution with CD25-depleted TBM spleen cells reversed the suppressive effect in both, TBM and naïve mice. Tumor-specific IFN-γ secretion was further enhanced when tumor targets expressed higher MHC class I and also MHC class II (D5CIITA vs. MCA-310CIITA) as elevated expression of both triggered more robust tumor-specific CD8⁺ and CD4⁺ cytokine production (by ICS, data not shown) and increased cytokine secretion (Figure 4A). Excitingly, T_E generated from lymphopenic, systemic tumor-bearing hosts reconstituted with CD25-depleted TBM spleen cells also exhibited the same level of therapeutic efficacy as shown for T_E generated in naïve lymphopenic hosts. These data indicate that CD25-depletion in the RLM generates a window of opportunity for effective vaccination in the absence of tumor-induced Treg even in the presence of progressive, systemic tumor growth.

Figure 4. IFN-γ secretion and therapeutic efficacy of T_E from naïve and irradiated TBM RLM reconstituted with CD25-depleted TBM spleens prior to vaccination.

(A) IFN-γ secretion by T_E from RLM. 2 × 10⁶ T_E generated in lymphopenic mice, reconstituted with naïve, TBM total and TBM CD25-depleted spleen cells were stimulated for 24h in CM only, or 10⁵ tumor cells as indicated. D5 and MCA-310 tumor cells were further transfected to express MHC-class II-transcriptional-factor (CIITA). Supernatants were collected and IFN-γ (pg/mL) secretion was determined by ELISA. Data show mean ± SEM of duplicates from two independent, pooled experiments. *p<0.05, two-tailed Student's t-test.

(B) T_E were generated from 8-day TVDLN cells after in vitro activation/expansion. TVDLN were harvested from wt500R lymphopenic naïve or 8-day tumor-bearing mice, reconstituted with 20 × 10⁶ total or in vitro CD25-depleted spleen cells from naïve or 8-day TBM mice and immediately vaccinated with D5-G6. Each group represents 5 wt normal animals with established 3-day D5 pulmonary metastases, controls received only 90 000 IU IL-2 i.p. for 4 consecutive days, all other groups shown received 30 × 10⁶ T_E from indicated RLM donors including IL-2 i.p. On day 14, lungs were harvested, fixed in Fekete's solution and pulmonary metastases were counted. Data are pooled from two independent experiments and show mean + SEM. *p<0.05, treatment groups and control; †p<0.05, between treatment groups using the Wilcoxon rank sum test.

Tumor-induced suppression prevents priming of tumor-specific, therapeutic effector T cells

Next, we investigated if vaccination of a lymphopenic host reconstituted with TBM T cells actually primed tumor-specific T cells, but they could not be detected because they were suppressed by a subpopulation of CD4⁺ Treg cells during in vitro activation and expansion or thereafter. To address this hypothesis, CD4⁺ T cells were selectively depleted, using magnetic beads, immediately after the isolation of TVDLN cells. Instead of depleting CD25⁺ cells, CD4 mAb were used to deplete all CD4⁺ (CD25⁺ and CD25^-) T cells, since activated CD4⁺ and CD8⁺ T cells would be expected to express CD25 and thus be depleted by anti-CD25 beads. Importantly, CD4⁺ T cells are required in vivo during priming (supplemental table 1). Using this strategy, we found that eliminating CD4⁺ T cells from freshly isolated TVDLN lead to the recovery of effector T cells with therapeutic efficacy if RLM were reconstituted with naïve spleen cells but not when RLM were reconstituted with TBM spleen cells (Figure 5A). The same pattern of therapeutic efficacy was observed when CD4⁺ T cells were depleted following in vitro activation and expansion (Figure 5B).

Figure 5. Effect of CD4-depletion on vaccine therapeutic efficacy.

(A) Depletion of CD4⁺ T cells from TBM TVDLN fails to restore therapeutic efficacy. TVDLN were harvested from wt500R lymphopenic mice reconstituted with either 20 × 10⁶ spleen cells from naïve mice or 8-day TBM and immediately vaccinated with D5-G6. Eight days later, TVDLN cells were harvested and either magnetically depleted of CD4⁺ T cells or un-manipulated and then activated in vitro and expanded. Equivalent 20 × 10⁶ CD8⁺ of the resulting T_E were then adoptively transferred into mice with 3-day D5 pulmonary metastases and lung metastases were enumerated 13 days later. (B) Depletion of CD4⁺ T cells from T_E generated after in vitro activation and expansion of TBM TVDLN cells fails to restore therapeutic efficacy. Mice were treated as described in A, with the exception that isolated T_E were expanded in vitro. Lung metastases were enumerated on day 14 after tumor inoculation. Data are from two independent, pooled experiments. *p<0.05, treatment vs. control groups; †p<0.05, between treatment groups; Wilcoxon rank sum test.

These results suggest that tumor-induced Treg prevent vaccine-induced priming of tumor-specific effector cells and suggests that successful intervention to recover therapeutic efficacy by vaccination in a host with progressive tumor growth can only be achieved if tumor-induced CD4⁺CD25⁺ Treg are reduced or eliminated prior to vaccination.

Unrelated tumor can suppress priming of D5 tumor-specific, therapeutic effector T cells

Finally, we asked whether the suppression observed in this system was specific for the D5 melanoma or whether established tumors of other histological types also produced immune suppression. To assess this question, spleen cells from mice bearing either MCA-310 sarcoma, 3LL lung carcinoma or MPR5 prostate carcinoma were used in the RLM model. Lymphopenic mice were reconstituted with spleen cells from either naïve, D5 TBM, MPR5 TBM, 3LL TBM or MCA-310 TBM, vaccinated with the D5-G6 melanoma vaccine and TVDLN were harvested to generate T_E. Strikingly, T_E generated from all animals reconstituted with spleen cells from TBM and vaccinated with D5-G6 failed to generate T_E that secreted D5-specific IFN-γ and either failed or exhibited drastically reduced therapeutic efficacy (Figure 6A and 6B). The reconstitution of lymphopenic hosts with CD25-depleted MPR5, 3LL or MCA-310 TBM spleen cells partially or completely restored the immune response to vaccination. Depletion of CD4+CD25+ T cells from MPR5, MCA-310 and 3LL tumor-bearing mouse donor T cells restored priming against a melanoma vaccine to generate D5 melanoma-specific effector T cell responses. The data clearly shows that CD4+CD25+ Treg induced by various histological different tumors can inhibit subsequent priming of effector T cells by histological distinct tumor vaccines. These data suggest that the generation of tumor-induced Treg may be antigen-non-specific.

Figure 6. IFN-γ secretion and therapeutic efficacy of T_E generated in lymphopenic, D5G6-vaccinated hosts is impaired when reconstituted with unrelated TBM donor spleen cells.

(A) IFN-γ secretion of T_E. 2 × 10⁶ T_E generated in lymphopenic mice, reconstituted with naïve, D5 TBM total, D5 TBM CD25-depleted spleen cells or spleen cells from MCA-310, 3LL or MPR5 TBM were stimulated for 24 h in CM only, or 1 × 10⁵ tumor cells as indicated. D5 and MCA-310 tumor cells were further transfected to express MHC-classII-transcriptional-factor (CIITA) to express MHC-class II and enhanced MHC-class I. Supernatants were collected and IFN-γ secretion was determined by ELISA. Data show mean ± SEM of duplicates from 2 independent, pooled experiments. *p<0.05, two-tailed Student's t-test.

(B) T_E generated in lymphopenic, D5G6-vaccinated hosts reconstituted with unrelated TBM donor spleen cells lack therapeutic efficacy against D5 melanoma. T_E were generated from 8-day TVDLN cells after in vitro activation/expansion. TVDLN were harvested from wt500R lymphopenic mice that were reconstituted with 20 × 10⁶ total or in vitro CD25-depleted D5 TBM spleen cells or MCA-310 TBM, 3LL TBM, or MPR5 TBM spleen cells and immediately vaccinated with D5-G6. Each group represents 5 wt normal animals with established 3-day D5 pulmonary metastases, controls received only 90 000 IU IL-2 i.p. for 4 consecutive days, all other groups shown received 30 × 10⁶ T_E from indicated RLM donors including IL-2 i.p. On day 14, lungs were harvested, fixed in Fekete's solution and pulmonary metastases were counted. Data represents 2 of 4 consecutive experiments and is displayed as mean ± SEM. *p<0.05, treated group vs. negative control (no AIT); †p<0.05, treated groups vs. positive control (Naïve total); Wilcoxon rank sum test.

Discussion

Over the past two decades, innovative vaccine strategies have been applied to patients with cancer, however objective clinical responses continue to be rare events. The inability to induce large numbers of tumor-specific therapeutic T cells affords an explanation for this failure.

Cancer vaccines are intended to initiate, boost and sustain a therapeutic anti-tumor immune response in patients, and although there is strong evidence that vaccines induce protective immunity in naïve tumor-free animals, there is little evidence that vaccines alone can induce a therapeutic immune response in animals with established tumor. In recent years combinations of immune intervention as well as combinations of an immune agent with chemotherapy, radiation or other biologics have provided reasons for renewed optimism in regard to active specific immunotherapy [26-30].

In support of the concept that a higher frequency of tumor-specific T cells are required in order to obtain objective tumor responses, Rosenberg and colleagues observed a significant correlation with objective clinical responses and maintenance of at least 5% circulating tumor-specific T cells for two weeks in non-myeloablative patients receiving adoptive T cell immunotherapy [27, 31]. While the results from adoptive immunotherapy trials, where large numbers of T cells were infused, might not fully correlate with the initial priming efficacy necessary for successful active, specific immunotherapy with cancer vaccines [32, 33], they provide proof of concept for what may be necessary to obtain in order to mediate tumor regression. A window of opportunity to successfully administer a tumor vaccine generated by the optimal conditioning of the recipient might allow for the generation of a substantially higher frequency of tumor-specific T cells.

Our novel approach exploited the observation that a tumor vaccine administered during homeostasis-driven expansion of T cells in a lymphopenic host dramatically increased the frequency and therapeutic efficacy of effector T cells [12, 28, 34]. However, similar to the studies noted above, the vaccinated naïve lymphopenic mice were reconstituted with immune cells from naïve mice free of any of the effects of tumor-induced suppression. Here, we evaluated whether this strategy would be effective, if T_E were derived from mice bearing advanced systemic tumor burden, a situation relevant to the clinical application of this research. The data show that reconstitution of lymphopenic hosts with spleen cells from mice bearing systemic tumor for eight to twelve days failed to generate IFN-γ secreting tumor-specific effector T cells from TVDLN (Figure 3C and data not shown). Since the presence of systemic tumor burden in the donor mouse impaired the generation of tumor-specific T cells we considered the possibility that tumor cells induced CD4⁺ Treg cells that were the mechanism for the loss of tumor-specific T cell function. We have previously reported that D5-G6, a GM-CSF secreting melanoma vaccine, effectively induced CD8⁺ tumor-specific T_E with therapeutic function in naïve mice deficient in CD4⁺ T cells [2]. Therefore, since CD4⁺ T cells were not necessary for priming tumor-specific, therapeutic TE when naïve, replete CD4-deficient mice were vaccinated in this model, we posited that depletion of CD4⁺ TBM spleen cells would eliminate Treg and reverse the failure to prime T_E. To our surprise, depletion of CD4⁺ T cells prior to reconstitution completely eliminated the ability to generate tumor-specific CD8⁺ T_E, regardless of whether naïve or TBM spleen cells were used for the reconstitution of lymphopenic hosts.

This is in striking contrast to data obtained in normal, intact MHC class II^-/- mice or wt mice depleted of CD4⁺ T cells by anti-CD4 mAb administration. Here, in the RLM model the presence of CD4⁺ T cells during reconstitution and exposure to the vaccine is essential for the priming of CD8⁺ T_E.

These observations provide insights into the impact of non-myeloablative lymphopenia conditioning regimens, that have a pronounced and long-term depleting effect on CD4⁺ T cell numbers [35-38], may have on the ability of the host to respond to vaccination. Our data would suggest that depletion of CD4⁺ T cells during an initial priming vaccination in lymphopenic patients could impair the development of a tumor-specific CD8⁺ T cell response. This is in spite of the fact that the CD8⁺ T cells would be undergoing homoeostasis-driven expansion. These results also suggest that when agents with pronounced long-term CD4-depleting effects are administered to patients, strategies to reconstitute this functional pool of CD4⁺ T cells should be considered prior to administering a cancer vaccine.

To preserve the majority of CD4⁺ T cells we targeted the regulatory T cell population by depleting CD25⁺ T cells. Strikingly, selective depletion of the CD25⁺ subset before reconstitution of the lymphopenic host reversed the profound immune deficit induced by B16BL6-D5. Furthermore, the restoration of suppression through adding back CD25⁺ TBM T cells to the CD25^- spleen cells confirms their role in blocking priming of T_E.

Since CD25-depleted TBM spleen cells were also effective when used to reconstitute lymphopenic mice with systemic tumor, the tumor-induced suppressive effect can at least be overcome transiently, allowing for the generation of therapeutic T_E. Essentially, this approach allows the immune system of a tumor-bearing host to be “re-booted” at the time of vaccination. It remains to be seen whether this one-time intervention will be sufficient to overcome the continuing influence of systemic tumor burden that will remain in patients.

The spleen cells from mice bearing three histologically different tumors and which lack known cross-reactive antigens, also contain CD25⁺ T cells that can block therapeutic efficacy. This suggests that antigen-independent mechanisms may play an important role in suppression mediated by tumor-induced Treg. There is no clear evidence in non-transgenic models for antigen-specific regulatory T cell function that inhibits priming of CD8+ effector T cells. This report is the first to show, that in a reconstituted, lymphopenic environment, tumor vaccination fails in the presence of Treg that were exposed to histological distinct tumors while naïve Treg do not show measureable interference in priming in this environment. However, we recognize the possibility that tumor-induced Treg generated in these tumor-bearing mice may recognize a previously unknown antigen shared by all four tumors. Alternatively, a toxic effect secondary to tumor burden might be responsible for the induction of suppression. In support of this not being a generalized non-specific effect of tumor burden, mice bearing MPR4 tumors with similar levels of tumor burden to those in Figure 6A/B were able to provide spleen cells for RLM studies that did generate tumor-specific T cells with therapeutic function in adoptive transfer studies (data not shown). While studies are continuing, one likely possibility is that heterogeneity in expression of regulatory T cell-inducing factors is responsible.

The question remains as to how tumor-induced Treg suppress the immune response. Currently, it is unclear, if different tumors trigger the generation of Tregs with different functions. Recent reports suggest at least 7 different ways for Treg to block T cell function: e.g. IL-10 [39], TGF-β [40, 41], PGE₂ [42], GITR [17, 43], blocking signaling through MHC class II [44], over-expression of CTLA-4 [45], CD39/CD73 signaling [46] or possible secretion of IL-35 [47]. There could be a common mechanism or groups of mechanisms triggered in regulatory CD4⁺ T cells by activated APC and/or T cells or for suppressing either antigen presentation/ maturation of APC and/or T_E activation. The answer to these questions will be vital to develop selective measures to manipulate tumor-induced Treg in the tumor microenvironment during priming of tumor-specific T_E to generate higher therapeutic efficacy. Future studies trying to identify key mechanisms tumors use to induce these Treg could serve as better targets to manipulate the induction or the maintenance of Treg. Possible intervention utilizing the inhibition of signaling, e.g. of the TGF-β receptors [51, in press], prostaglandin pathways, or further in-vivo manipulation of CD4⁺ T cell subsets are under consideration to further enhance the priming effector and effector/memory T cells via tumor vaccination.

The results described here strongly suggest that effective vaccination against tumor in a recipient with progressive tumor burden will require the elimination of Treg. Based on these results a translational clinical trial using this approach has been initiated in patients with melanoma.

Materials & Methods

Animals

Female C57BL/6J (wt) and RAG-1 KO (B6.129S7-Rag1^tm1Mom) breeders were purchased from Jackson Laboratory (Bar Harbor, ME). Breeder colonies were kept in pathogen-free conditions and mice of age 6-14 weeks were used for experiments. Mice were treated according to the principles of the Guide for the Care and Use of Laboratory Animals, National Research Council, 1996. All animal research protocols were approved by the Earle A. Chiles Research Institute animal care and use committee.

Tumor cell lines

B16BL6-D5 (D5) is a poorly immunogenic subclone of the spontaneously arising B16BL6 melanoma derived from C57BL/6 mice that expresses low levels of MHC class I (K^b) and no class II (I-A^b). D5-G6 is a stable clone of D5 that was transduced with a murine GM-CSF retroviral MFG vector (provided by Dr. M. Arca, University of Michigan, Ann Arbor, MI). D5-G6 cells secrete approximately 200 ng/ml/10⁶ cells/24 hours GM-CSF. MCA-310 is a methyl-cholanthrene-induced, weakly immunogenic, fibrosarcoma of C57BL/6 mice. D5^CIITA and MCA-310^CIITA are stable cell lines of D5 and MCA-310 that have been transduced with the transcriptional factor CIITA to express MHC class II and increased levels of MHC class I. 3LL (Lewis lung carcinoma) and MPR5 (murine prostate carcinoma) were also used to establish systemic metastases in TBM mice.

Culture conditions

Lymphocytes and tumor cells were cultured in complete medium (CM), which consisted of RPMI 1640 supplemented with 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 2 mM L-glutamine, 50 μg/ml of gentamicin sulfate (all from Cambrex-BioWhittaker, now Lonza, Walkersville, MD.), 50 μM 2-mercaptoethanol (Sigma, St. Louis, MO, USA.), and 10% fetal bovine serum (FBS, premium quality, US origin only) (Cambrex-BioWhittaker, now Lonza). Tumor cells were harvested 2 to 3 times per week by brief trypsinization (Trypsin, BioWhittaker, now Lonza) and maintained in T-150 or T-225 culture flasks for no longer than 14 days. FACS buffer contained 2% FBS in PBS.

Lymphopenia, Magnetic Cell Depletion, Reconstitution and Generation of T_E and Adoptive Immunotherapy

20 × 10⁶ fresh spleen cells obtained from either naïve mice or mice carrying 8-14 days systemic D5 tumor (TBM: established by tail vein injection of 10⁵ D5 tumor cells in HBSS) were used to reconstitute lymphopenic hosts, which were either 500R irradiated wt (wt500R) or RAG-KO mice. For in vivo depletion of CD4⁺ T cells, naïve or TBM mice received i.p. injection of anti-CD4 mAb (1:40 ascites from hybridoma GK1.5 (TIB207), ATCC, Manassas, VA) on day -3 and -1 before spleen cells were harvested. For in vitro depletion of CD4⁺ and CD25⁺ cells, spleen cells were first stained with either anti-CD4^PE or anti-CD25^PE mAb (Pharmingen, San Diego, CA) for 20 min. at 4°C in HBSS, washed, stained with anti-PE-Magnetic Beads (Miltenyi Biotec, Auburn, CA) for 20 min. at 4°C, washed and magnetically separated using a Vario MACS™ (Miltenyi). Total, CD4-depleted or CD25-depleted spleen cells were used to reconstitute lymphopenic mice. In indicated experiments, 1 × 10⁶ CD25⁺-enriched spleen cells from tumor-bearing mice were “added back” to the CD25-depleted naïve and tumor-bearing spleen cell populations prior to reconstitution with a total of 20 × 10⁶ spleen cells.

Reconstituted, lymphopenic mice (RLM) were immediately vaccinated with 10⁶ D5-G6 tumor cells injected subcutaneously in all four flanks. In some experiments, normal, non-RLM wt mice were vaccinated as controls. Eight days following tumor inoculation the tumor-vaccine draining lymph nodes (TVDLN) were harvested, washed, resuspended at 2×10⁶ cells per ml in CM and activated in 24-well plates with anti-CD3 and anti-CD28 as described previously [12]. After two days of activation, the T cells were harvested and expanded in CM containing 60 IU rhIL-2/ml (kind gift from Chiron, Emeryville, CA) for three days. T cells were then harvested, washed twice in HBSS and used in cytotoxicity, intracellular cytokine staining, cytokine release assays and adoptive transfer studies. For adoptive transfer experiments, wt mice bearing 3-day established pulmonary metastases were generated by i.v. inoculation with 0.2 × 10⁶ D5 tumor cells. Effector T cells were transferred i.v. and recipient mice received 90,000 IU IL-2 i.p. q.d. for 4 days. In some experiments, mice bearing 3-day D5 pulmonary metastases were irradiated (500R) just prior to adoptive effector T cell transfer. Mice were sacrificed 14 days following tumor inoculation, lungs were resected, fixed in Fekete's solution and the number of pulmonary metastases enumerated. In all experiments T_E used for in vivo studies were analyzed in parallel for their in vitro cytolytic capability and/or cytokine profile.

ELISA

T_E from RLM and wt mice were washed, resuspended in CM and cultured at 2 × 10⁶ cells/well in a 48-well plate. The cells were either left unstimulated (negative control) or stimulated with 10⁵ tumor cells (D5, MCA-310), or anti-CD3 (positive control). Supernatants were harvested after 20-24 hours and assayed for IFN-γ, IL-5 and IL-10 using commercially available reagents (Pharmingen, San Diego, CA).

In vitro cytotoxicity assay

For ⁵¹Cr-release assays, D5 or MCA-310 were incubated with 100 μCi Na⁵¹CrO₄ for 90 min., washed and plated into 96-well round-bottom plates with 10⁴ tumor target cells/well. Target cells were incubated with T_E in triplicate at the indicated E:T ratios in 200μl of CM at 37°C in a CO₂ incubator. The supernatant was harvested after 6h, counted and the mean percent lysis determined as previously described.

Intracellular cytokine and polychromatic flow cytometric analysis

T_E from RLM and wt mice were stimulated for 12 hours in the presence of 5μg/ml Brefeldin A (Sigma) in CM only (no stimulation), with specific tumor (D5), unrelated syngeneic tumor (MCA-310) or immobilized anti-CD3. T_E were harvested and stained with anti-CD8^FITC and anti-CD3^CyChrome™ mAb, fixed and permeabilized in Cytofix/Cytoperm™ and stained intracellularly with anti-TNF-α^PE or anti-IFN-γ^PE mAb (all from Pharmingen). Analysis was performed on 50,000 gated CD8⁺CD3⁺ T_E with a FACS™ Calibur and Cellquest software (Becton & Dickinson, San Diego, CA, USA). Data are presented as the percentage of TNF-α⁺ or IFN-γ⁺ cells from total CD8⁺CD3⁺ T_E. For 8-color / 10-parameter flow cytometry analysis fresh spleen cells from wt and 8-day systemic D5 tumor-bearing mice were harvested, Fcγ receptors were blocked (2.4G2 rat mAb; Pharmingen) for 20 min at 4°C, washed in FACS buffer and stained in multiple 20 min steps with anti-CD3^PE-Cy7, anti-CD4^APC-Cy7, anti-CD25^PE, anti-CD62L^APC, anti-CD44^Cychrome, anti-CD69^FITC (Pharmingen), anti-CD8^PE-TR (CALTAG Lab., Burlingame, CA, USA) mAb and with rhu ELC:Fc (CCL19:Fc; ligand to CCR7; kind gift from Ullrich v. Andrian, Dept. of Pathology, Harvard Medical School [48]), followed by biotinylated anti-human IgG-Fcγ-specific mAb and Streptavidin-labeled PE (Jackson Immunoresearch Lab., West Grove, PA, USA). Controls were stained with anti-CD3^PE-Cy7 only, or one other listed surface marker. A 10-parameter flow cytometric analysis on Forward and Side Scatter (FCC/SCC) adjusted, CD3⁺ cells was performed using a MoFlo cell sorter (Cytomation, Fort Collins, CO) and multiple vector analysis was evaluated on spleen cells from tumor-bearing as compared to naïve animals. Flow histograms show a representative phenotype of tumor-bearing spleen cells and the mean ± SEM of the percentage of CD3⁺CD4⁺CD25⁺ over all CD3⁺CD4⁺ T cells from 3 independent experiments.

Statistical Analysis

Statistical analysis of the number of pulmonary metastases was performed by Wilcoxon rank sum test; and a two-tailed t-test using Graph Pad Prism 5.0 (Graph Pad Software Inc.) was performed to analyze the differences in cytokine secretion and the percentage of IFN-γ secreting T cells.

Abbreviations

AIT

adoptive Immunotherapy

D5

poorly immunogenic subclone of B16 melanoma tumor cell line

D5-G6

stable GM-CSF-transduced clone of D5 melanoma

MCA-310

methyl-cholantherene-induced, poorly immunogenic fibrosarcoma

RLM

reconstituted, lymphopenic mouse

TBM

progressive tumor-bearing mouse

T_E

effector T cells

Treg

regulatory T cells, CD4+CD25+ phenotype

TVDLN

tumor-vaccine draining lymph node