Background
IL-21 has potent immunomodulatory effects on T cells and NK cells [
1]. However, the current understanding of the effects of IL-21 on the regulation of both antigen-independent homeostatic and antigen-stimulated proliferation and activation of naive and memory CD8
+ T cells is confounded by conflicting data. IL-21 has been reported both to synergize with IL-15 to increase homeostatic antigen-independent proliferation of naive and memory CD8
+ T cells [
2,
3], and to inhibit IL-15 induced homeostatic proliferation of memory T cells [
4]. While evidence suggests IL-21 enhances primary antigen-stimulated T-cell proliferation and functional activation [
3‐
6], there is disagreement over whether this effect occurs during priming [
5], or during the late expansion phase of antigen-specific proliferation [
6]. Recent data suggests IL-21 primarily increases the frequency and absolute number of in vitro stimulated antigen-specific CD8
+ T cells by enhancing the rate of proliferation rather than by reducing apoptosis [
5]. Alternatively, other studies indicate IL-21 maintains increased numbers of antigen-stimulated effector and long-term memory CD8
+ T cells by reducing apoptosis while maintaining a low rate of cell division [
6]. These conflicting data may be attributed to differences in the in vitro and in vivo models employed in the individual studies. However, they suggest that much work remains in the effort to elucidate the mechanism of IL-21 regulation of T-cell immune responses, including any potential role it may play in the regulation of T-cell mediated anti-tumor immunity through the induction of functionally active tumor-specific effector and memory T cells (Reviewed in [
1]).
In vivo data demonstrate that IL-21 enhances the anti-tumor function of T-cells and NK cells. Several studies have shown that genetically engineered IL-21-secreting murine tumors activate potent NK and CD8
+ T cell mediated anti-tumor responses [
7‐
9]. IL-21 was also compared to IL-2 and IL-15 injected intraperitoneally (IP) subsequent to challenge with the ova-expressing E.G7 thymoma in either naive C57BL/6 mice, or in mice infused with OT-1 transgenic CD8
+ T cells [
6]. In both models, IL-2 and IL-15 delayed tumor growth, but only IL-21 resulted in significant prevention of tumor progression and improved survival beyond 50 days from tumor challenge. The combination of IL-21 and IL-15 was also administered IP subsequent to human gp100
25–33 (hgp100
25–33) peptide vaccination of lymphopenic, tumor-bearing (B16F10 melanoma) C57BL/6 mice [
3]. Before vaccination mice received in vitro stimulated (IVS) pmel transgenic (Tg) CD8
+ T cells, which have a T-cell receptor specific for the murine H-2D
b restricted gp100
25–33 peptide of the B16F10 melanoma associated gp100 protein, and strongly cross react with hgp100
25–33 [
10]. The combination of IL-21 + IL-15 resulted in prolonged tumor regression and survival out to 32 days. Mice treated with the vaccine plus IL-15 or IL-21 alone all died of tumor within 32 days of treatment. No data was presented demonstrating long-term survival (>60 days) of mice treated with IL-21 + IL-15. Notably, IL-2 was not tested in vivo in combination with IL-21 in this study primarily because the combination of these two cytokines failed to drive antigen-independent homeostatic proliferation of murine CD8
+ splenocytes in vitro [
3], and the concern that high-dose IL-2 has been shown to decrease memory CD8
+ T-cell function by inducing regulatory T-cells and increasing antigen-driven apoptosis [
11‐
14]. However, other data have demonstrated that the in vitro maintenance of antigen-stimulated murine T cells in low-dose IL-2 resulted in the proliferative expansion of cells with a memory phenotype similar to that induced by IL-15; these cells were capable of long-term survival in vivo, and potent proliferative and functional responses after secondary in vivo challenge [
15]. Thus, treatment with IL-21 + low-dose IL-2 after the adoptive transfer of pmel Tg CD8
+ T cells into mice with established B16 melanoma tumors (pmel Tg/B16 model) might yield anti-tumor effects comparable to or greater than those observed with IL-21+IL-15.
The pmel Tg/B16 melanoma model has been employed repeatedly in recent studies to model immunotherapeutic strategies for treatment of pre-existing disease. Most of these experiments have focused on the adoptive transfer of antigen-educated pmel CD8
+ splenocytes stimulated in vitro with hgp100
25–33 peptide and IL-2, and the subsequent in vivo expansion of these highly activated cells in tumor-bearing mice with hgp100
25–33 vaccination and follow-on cytokine therapy [
3,
16‐
18]. While this strategy has resulted in significant tumor regression, little data has been presented demonstrating long-term curative effects (≥ 100 days) with IL-2 alone [
16,
17], IL-15 alone [
18], or the combination of IL-15 and IL-21 [
3] – except in mice with small tumors (≤ 10 mm
2) at the time of treatment, or after the adoptive transfer of very large numbers (10
7) of IVS pmel CD8
+ T cells [
16]. The adoptive transfer of highly activated IVS T-cells may not produce long-term anti-tumor curative effects due to functional and proliferative exhaustion [
19‐
21]. This conclusion is supported by a series of provocative experiments using the pmel Tg/B16 model in which the adoptive transfer of naive or "early" IVS-effector pmel CD8
+ T cells into lymphopenic tumor-bearing recipient mice resulted in much more durable tumor regression following hgp100
25–33 vaccination and IL-2 therapy compared to the poor therapeutic effects observed using IVS expanded, highly stimulated intermediate or late-stage effector T cells [
22]. Thus, as described herein, the adoptive transfer of naïve, tumor antigen-specific CD8
+ T cells in this model prior to vaccine and IL-21 + low-dose IL-2 therapy may provide a better opportunity to study methods of generating both effector and long-term anti-tumor memory T-cell function than can be achieved with highly activated IVS T cells.
There have been few published studies describing the anti-tumor effects of recombinant IL-21 in conjunction with hgp10025–33 vaccination in B16F10 tumor-bearing lymphopenic mice receiving naive pmel Tg CD8+ T cells. There have, to our knowledge, been no reports describing the potential synergy of IL-21 + low-dose IL-2 in this or other tumor models. Herein we describe the long-term curative effect of combined IL-21 + low-dose IL-2 cytokine therapy in lymphopenic, tumor-bearing C57BL/6 mice infused with naive pmel splenocytes and vaccinated with the hgp10025–33 melanoma peptide.
Discussion and conclusions
We report for the first time the anti-tumor effects of a treatment strategy that combines lymphopenic conditioning, adoptive transfer of naive antigen-specific CD8
+ T cells, tumor antigen-specific vaccination, and cytokine treatment with IL-21 + low-dose IL-2. Using a well established murine model of pre-existing disease, the pmel Tg/B16 model, our results show that combined IL-21 + low-dose IL-2 therapy delayed B16F10 melanoma growth in the majority of mice, and resulted in a significant increase in tumor-free survival (46%) out to ≥ 150 days compared to mice treated with IL-21 or IL-2 alone. Surviving mice exhibited tumor-specific protective immunity since they resisted B16F10 rechallenge but succumbed to the unrelated 3LL tumor. These data are in contrast to other studies in the pmel Tg/B16 model, which employed the adoptive transfer of highly activated IVS pmel CD8
+ T cells (stimulated with hgp100
25–33 peptide and IL-2 or IL-15) rather than naive pmel T cells [
3,
17,
18]. After lymphopenic conditioning and T-cell transfer to tumor-bearing mice, therapy was administered in these studies using different combinations of hgp100
25–33 vaccination and cytokine treatment. Thus, Lou et al [
17] vaccinated with hgp100
25–33 peptide-pulsed DCs, and administered high-dose IL-2 (1.2 × 10
6 IU/day) for three days beginning the day of vaccination. Klebanoff, et al [
18] vaccinated with a recombinant fowl pox virus encoding hgp100
25–33 (rFPhgp100) after adoptive transfer of IVS pmel CD8
+ T cells grown in IL-2 or in IL-15 cultures. Follow-on cytokine therapy consisted of high-dose IL-2 (1.2 × 10
6 IU/day) for three days. In the only published study using IL-21 in the pmel Tg/B16 model Zeng and coworkers [
3] also vaccinated with rFPhgp100
25–33 following adoptive transfer of IL-2 cultured pmel T cells, and animals were treated over three days with IL-15 (20 μg/day), IL-21 (20 μg/day), or both cytokines at the same doses. None of these treatment strategies resulted in long-term inhibition of tumor growth beyond 30–35 days after treatment unless very large numbers (4–6 × 10
6) of IVS pmel CD8
+ T cells were adoptively transferred [
17]. Furthermore, in contrast to our results, none of these studies presented data describing long-term tumor-free survival ≥ 150 days with any combination of transferred T cells, vaccine administration and cytokine therapy. Another group compared the ability of IL-2, IL-15 or IL-21 to augment anti-tumor immunity in C57BL/6 mice challenged with the OVA-expressing E.G7 thymoma [
6]. Cytokine therapy was initiated 48 hours after tumor inoculation – well before the establishment of vascularized tumors, and was administered every other day for two weeks thereafter. In this model IL-21 (20 μg/day) did result in prolonged tumor-free survival out to 100 days in approximately 25–30% of treated mice. Neither low-dose IL-2 (2 × 10
3 IU and 20 × 10
3 IU/day) nor IL-15 (5 μg and 50 μg/day) produced a survival frequency that was as high. All animals that survived beyond 100 days were protected from rechallenge with E.G7 tumor cells. Importantly, this report also showed that the 2-week cytokine administration could begin as late as day 12 following tumor inoculation and still produce a therapeutic response [
6]. This study also demonstrated the ability of IL-21 to augment immunity to the OVA-expressing E.G7 tumor in the absence of previously antigen-educated CD8
+ T cells. However, the experimental design did not test the anti-tumor effects of IL-21 in a model of established bulky disease such as the pmel Tg/B16 system, in which the cognate tumor antigen is a relatively weak self-tumor antigen rather than a strongly immunogenic foreign protein. To date only the combined therapeutic effects of IL-21+ low-dose IL-2 reported here have resulted in long-term survival of tumor-free mice in the pmel Tg/B16 model. As noted, in most of the previous studies using the pmel Tg/B16 melanoma model very high doses (1.2 × 10
6 IU/day) of IL-2 were used in combination with the adoptive transfer of large numbers (up to 10
7) of highly activated IVS pmel CD8
+ T cells [
16‐
18] However, treatment was initiated after very large, bulky (50–100 mm
2) tumors were established. By contrast, cytokine therapy was initiated in our experiments when vascularized tumors were smaller (26 mm
2). Long-term survival using our current IL-21 + IL-2 therapeutic model may decrease with larger established tumors; experiments designed to test and optimize IL-21 +low-dose IL-2 therapy in mice with larger vascularized tumors are ongoing.
The H-2D
b-restricted NV peptide was used in all experiments as a putative CD4 "helper" antigen source [
23,
24] However, our recent experiments suggest NV peptide vaccination produces only a modest increase in circulating pmel CD8
+ T cell numbers over that produced by hgp100
25–33 alone (unpublished data – H.M. Hu, EACRI). CD4 helper T cell function in our DC-based vaccine system may have been provided by fetal calf serum (FCS) proteins/peptides associated with the DCs in the vaccine – both DCs and the B16F10 melanoma cells used in the tumor inoculum were cultured in medium containing FCS. The potential for FCS antigen-induced CD4 T cell activation is also suggested by other hgp100
25–33-pulsed DC vaccine studies in the pmel Tg/B16 model in which anti-tumor therapeutic effect was achieved in the absence of any known source of CD4-specific antigen [
17].
There is limited information on IL-21-mediated changes in the cell surface and functional phenotype of CD8
+ T-cells associated with tumor immunity. In this report, we show that IL-21 alone and IL-21+ low-dose IL-2 increased the in vivo frequency of Tcm T cells as defined by concomitant CD62L
+/CD127
+ staining[
29], compared to the vaccine only control or to cells from IL-2-treated mice (Figure
5). IL-21 + IL-2 treatment also sustained the highest percentage of Tem (CD62L
-/CD127
+) pmel CD8
+ T cells [
29] compared to any other test group. This observation correlated with the result that IL-21+ IL-2 therapy also produced the highest absolute number of circulating pmel CD8
+ T cells during the beginning (day 14), peak (day 21) and end (day 28) of the hgp100
25–33 induced expansion of pmel CD8
+ T cells (Figure
4). As a consequence, IL-21 + IL-2 therapy produced the highest absolute number of circulating pmel CD8
+ Tem cells at each time point of the expansion and early contraction phases of the anti-hgp100
25–33 immune response, and also resulted in the highest absolute number of Tcm T cells in the peripheral blood (Figure
6). Although Tem absolute numbers decreased from day 14 to day 28, and Tcm numbers increased over this same period the data do not indicate if this is attributable in anyway to innate differences in IL-21 + IL-2 induced Tem vs Tcm proliferation. Future in vitro and in vivo studies will examine the relative proliferative potential of cognate antigen-driven purified populations of pmel CD8
+ Tcm and Tem cells. Notably, the tumor-specific protective memory response observed in long-term survivor mice was also associated with high frequencies of Tcm in the spleen and lymph nodes. Polychromatic (8 color) flow cytometry analysis of pmel CD8
+ splenocytes from IL-21 + IL-2 treated long-term survivor mice (> 150 days) indicated approximately 10%–12% of all CD8
+ T cells in the spleen and lymph nodes were pmel CD8
+ T cells, and 43% and 70% of these T cells in the lymph nodes and spleen respectively expressed a Tcm phenotype (CD62L
+/CD127
+/CD27
+/CD28
+)(unpublished data). Recent data have demonstrated that enriched pmel CD8
+ Tcm T cells conferred a potent in vivo anti-tumor recall response upon adoptive transfer – leading to the eradication of large established tumors in the pmel Tg/B16 model [
35]. Thus, as described herein, IL-21 + IL-2 therapy favored the in vivo expansion and maintenance of such tumor antigen-specific Tcm T cells from naive pmel CD8
+ T cell precursors which were capable of providing long-term protective immunity. Preliminary phenotype analysis of IL-21 + low-dose IL-2 and hgp100
25–33 IVS pmel splenocytes similarly indicates the combination of both cytokines results in the proliferation of pmel Tg CD8
+ T cells with a Tcm (CD62L
+/CD127
+/CD27
+/CD28
+) phenotype (unpublished data – EACRI). This further suggests that IL-21 + low-dose IL-2 might be effective in supporting the cognate antigen-driven in vitro expansion of rare self-tumor antigen Tcm CD8
+ T cells from autologous PBMCs – perhaps in concert with procedures to remove tumor antigen-specific regulatory T cell effects. In addition to increasing tumor-antigen directed expansion and maintenance of Tcm CD8
+ T cells, data presented here demonstrates that IL-21 + IL-2 therapy also stimulated the expression of anti-tumor memory CD8
+ T cells with a Tc1 cytokine functional phenotype in mice with regressing tumors. By day 42 post-tumor inoculation IL-21 + IL-2 treated tumor-regressor mice had much higher absolute numbers of IL-2
+, IFNγ
+ and TNF-α
+ pmel CD8
+ splenocytes than IL-21 + IL-2 treated tumor-progressor mice with rapidly growing tumors (Figure
7). The relative absence of CD8
+ splenocytes with a Tc1 cytokine functional phenotype in tumor progressor mice may be attributable in part to a simple lack of sustained B16F10 tumor antigen stimulation. Previous published data from our institute using the pmel Tg/B16 model has demonstrated loss of MHC class I and gp100 expression in a high percentage of tumors in experimental animals with progressive disease due to immunoediting[
36]. Overall our data suggest that IL-21 + low-dose IL-2-mediated tumor regression and long-term survival was associated with elevated absolute numbers of tumor antigen-specific Tem and Tcm CD8
+ T cells with a Tc1 functional phenotype.
IL-21 + low-dose IL-2 induction of increased in vivo expression of Tc1 cytokine
+ Tcm and Tem CD8
+ T cells is a central observation for any putative mechanistic explanation of the anti-tumor therapeutic synergy produced using both cytokines. This observation examined in the context of recent studies describing the broad inhibitory effects of IL-21 on B lymphocytes, NK cells and DCs (Reviewed in [
37]) suggests a possible mechanism for the IL-21+ low-dose IL-2 induced expansion of Tcm CD8
+ T cells similar to what may occur during the late contraction phase of an ongoing immune response. Thus, CD4 T cell (Th2)-derived IL-21[
38] inhibits anti-IgM and IL-4-mediated B cell proliferation [
2] and enhances apoptosis of activated B cells [
39,
40], inhibits DC maturation and the ability of DCs to prime T cells [
41,
42]; and increases NK maturation and cytolytic function while inhibiting NK proliferation [
43]. Similarly, while IL-21 has generally been described as an inducer of antigen stimulated IFNγ and CTL function in CD8
+ T cells [
4,
8,
44,
45], our unpublished data in both murine and human in vitro experiments indicate IL-21 alone does not drive high levels of cognate antigen stimulated CTL proliferation. In vivo data presented here suggest that IL-21 works best in concert with low-dose IL-2 to support both optimal anti-tumor effector CTL function and expansion of CD8
+ memory T cells. Paradoxically, IL-21 + low-dose IL-2 treatment produced unexpectedly low frequencies of circulating effector pmel CD8
+ T cells (Figure
4) – perhaps due to cell trafficking to the primary tumor and sites of metastatic disease. Ongoing experiments are directed at the in situ analysis of the phenotype, cell number, and functional properties of tumor invasive pmel CD8
+ T cells in IL-21 + low dose IL-2 treated mice. Such studies should shed light on whether combined cytokine therapy induces trafficking of increased numbers of antigen-specific cytolytic CD8
+ T cells to the tumor site. Taken together, these observations suggest a dual role for IL-21 in dampening ongoing innate and adaptive immune responses (as might occur during the contraction phase of antigen-specific immunity), while concomitantly augmenting the expansion of antigen-driven long-term memory T cells. Recently published data demonstrate IL-21 treated cultures of antigen-stimulated cells produce CD8
+ T cells with increased TCR binding affinity [
5]. Other data show the IL-21 receptor (IL-21R) is upregulated upon TCR engagement [
46,
47]. Thus, Th2-derived IL-21 may "rescue" memory T cells by blocking further antigen-driven effector T cell differentiation through presently unknown mechanisms, while simultaneously increasing antigen binding affinity. The resulting increase in binding of low-levels of residual antigen (present during the contraction phase) could increase IL-21R expression and up-regulate increased IL-21 binding to memory T cells – thus further skewing the shift to memory T cells. Memory CD8
+ T cells in turn express amplified levels of CD122 [
48], which may facilitate increased high affinity IL-2R formation if other IL-2R chains (CD25 and CD132) are available, and potentially increase binding of the low levels of IL-2 which may be present during immunological contraction. The combination of the "arrested" Tcm memory phenotype (maintained by IL-21), and low dose IL-2-induced proliferation of these cells could result in increased Tcm cell numbers. Thus, IL-21 + low-dose IL-2 therapy may favor the enhanced expansion of long-term memory (Tcm) CD8
+ T cells through augmented memory T cell "rescue" and expansion mechanisms similar to those which may normally be present during the late contraction phase of an ongoing immune response. This concept is supported by the results of our study and other reports [
6,
8], which suggest IL-21 induced tumor immunity is most effective when IL-21 is administered several days (4–12 days) after initial tumor-antigen activation of T cells. The late acting anti-tumor effects of IL-21, and, in our study, IL-21 + IL-2 therapy suggests IL-21 may be acting directly on tumor antigen educated CD8
+ T cells rather than modulating early APC and/or T cell function associated with primary tumor antigen stimulation. Preliminary in vitro studies with purified naïve and hgp100
25–33 stimulated pmel CD8
+ T cells support this conclusion (data not shown).
In summary, the results of this study suggest that IL-21 + low-dose IL-2 cytokine therapy may "rescue" and augment proliferation of tumor antigen-specific memory CD8+ T cells, and thus provide an important new strategic component for effective cancer immunotherapy. The combined use of both cytokines may be most effective in the context of lymphopenic conditioning, tumor antigen-specific vaccination, and the adoptive transfer of autologous tumor-specific Tcm or "early" effector T cells. Ongoing work at our institute continues to focus on developing procedures for the IL-21 + low-dose IL-2-directed in vivo and in vitro expansion and maintenance of such Tcm and "early" effector subpopulations from small numbers of normally tolerized autologous self-tumor antigen-specific precursors.
Authors' contributions
HH designed the in vivo experiments and worked with PW and DH to perform the flow cytometry analysis of PBMC and splenocyte samples. GY assisted in all aspects of tumor inoculation, lymphopenic conditioning, tumor measurement and tissue harvest. HMH, CHC, and AOH provided key reagents and assisted in the design of controls which utilized these reagents in the experiments. WM was directly involved in drafting and revising the manuscript, and WGA performed all the statistical analysis. EBW designed and supervised all aspects of the experimental strategy and acquisition of data, and, with BAF and WJU, was involved in data interpretation and the critical review and revision of the manuscript.