We previously reported that high avidity CD8+ T cell populations specific for the immunodominant neu-derived epitope, RNEU420–429, are recruited to kill neu-expressing tumors following treatment with a neu-expressing vaccine and low dose Cy as a method to deplete Tregs. Here we report three new findings that for the first time elucidate the mechanism by which Tregs suppress high avidity T cell activation and migration in the immune tolerant environment of neu-N mice. First, high avidity cancer antigen-targeted T cells can effectively eradicate developing tumors in tolerant hosts when efficient numbers of activated T cells are available. However, activation and persistence of efficient numbers of high avidity T cells is only possible following reduction of Treg numbers. Importantly, we show for the first time that Tregs appear to be the major regulator of these high avidity effector T cells. Second, we also show for the first time that antigen-targeted low avidity T cells are unable to persist long-term and reach a necessary threshold of activation in tolerant hosts, even with Treg reduction. Third, a Cy-sensitive Foxp3+CD25low effector/memory Treg sub-population present in tolerant mice is the primary mediator of antigen-targeted high avidity T cell suppression within the tumor microenvironment.
Ideal vaccines should induce antigen-specific high avidity T cells that can migrate to sites of inflammation and express multiple effector cytokines 
. In practice, vaccines do induce high avidity T cells specific for infectious and cancer-associated antigens when administered prior to antigen exposure. However, there is an ongoing debate in the literature as to whether cancer antigen-specific high avidity T cells exist in the periphery of cancer-bearing hosts because many of these antigens are altered self proteins. Some studies have suggested that high avidity T cells are deleted, leaving only low avidity T cells for vaccine activation 
. Other studies have shown that high avidity antigen-targeted T cells exist, but are tolerized and unable to mediate tumor clearance 
. We previously reported that Cy mediated Treg depletion given in sequence with a neu-targeted vaccine induces endogenous high avidity T cells that mediate tumor clearance in 20–30% of mice 
. These prior studies could not address whether the low number of treatment successes was due to the existence of yet undefined tolerance mechanisms or the inability to induce large enough numbers of antigen-specific high avidity T cells. Using the transgenic T cells described here to model endogenous T cells, we demonstrate that larger numbers of vaccine-activated high avidity T cells are necessary to treat developing tumors in tolerized mice. Furthermore, these data suggest that vaccination alone is not enough to activate the number of endogenous T cells required for tumor clearance, and that a significant reduction in the numbers of a specific subset of Tregs is required to allow efficient vaccine activation of high avidity T cells. Thus, these data show that Tregs are indeed a main mechanism by which high avidity tumor antigen-specific T cell function is suppressed.
In addition to inducing high avidity T cells, an ideal vaccine should also induce a sufficient memory response to fight off future antigenic challenges 
. Our studies also show that approximately 60% of mice given Cy, vaccine, and adoptive transfer of high avidity T cells can resist a second tumor challenge given 3 weeks post adoptive transfer, indicating that our vaccine regimen can induce a partial memory type response. At this time point, Thy1.2+
high avidity T cells comprise a detectable population of the CD8+
T cells with effector and central memory phenotypes in neu
-N mice that wanes at 5 weeks (Figure S2
, and Data not shown). The partial response may be due to the re-establishment of tolerance in this system either by the re-establishment of Treg numbers, as we have shown, or by an as yet undetermined mechanism 
. Future experiments will address the longevity of the high avidity response in this system. However, we hesitate to make definitive conclusions regarding the memory phenotype of our adoptively transferred T cells, as others have clearly demonstrated that adoptively transferring high numbers of T cells alters the effector memory/central memory development ratio away from the endogenous situation 
Partial Treg depletion using low dose Cy plus a neu-targeted vaccine allowed for enhanced persistence, activation, and tumor-trafficking of adoptively transferred high avidity T cells, leading to tumor clearance in tolerant neu
-N mice. Thus, the availability of sufficient high avidity T cells alone is likely not enough to treat tumors in tolerant hosts. Instead, Cy treatment, either through Treg reduction or another mechanism, facilitates efficient high avidity T cell activation, persistence, and tumor trafficking. Studies have shown that lymph node residing T cells upregulate VLA-4, VLA-6, and CXCR3 in order to migrate to CXCL9 producing tumors 
. In our model, high avidity T cells transferred into Cy-treated neu
-N mice upregulated these trafficking receptors after adoptive transfer, allowing for tumor-infiltration, and tumor clearance by one week after Cy administration, which is the same time-point when depleted Treg populations begin to repopulate the tolerant host 
, and the transferred high avidity T cells lose IFNγ production. Thus, Cy-mediated Treg depletion allows high avidity T cells to upregulate the migration and activation signals necessary for tumor trafficking and anti-tumor activity. Until methods are available to deplete Tregs for longer than 1–2 weeks, it is not possible to determine whether vaccine induced endogenous T cells are capable of reaching the numbers achieved with adoptively transferred high avidity T cells that are required for tumor control in the majority of mice.
Although adoptively transferred high avidity T cells are efficiently activated to clear tumor in tolerant mice, their cytokine secretion capacity appears limited to IFNγ secretion. In contrast, these same high avidity T cells produce multiple cytokines (IFNγ, IL-2, and TNFα) when transferred into tumor-bearing non-tolerant FVB/N mice. Other groups have suggested that TNFα expression correlates with better effector function, cell persistence 
, and clinical outcomes 
. Thus, these data imply that a reduction in Treg numbers alone is not sufficient to optimally activate high avidity T cells. It is likely that fewer numbers of activated high avidity T cells would be required to clear the same tumor burden in tolerant neu
-N mice if these T cells were optimally activated. In support of this concept, the activation state of the high avidity T cells has recently been implicated in their ability to promote tumor clearance. Specifically, central memory T cells have been described as more “fit” to mediate tumor clearance due to their less-differentiated state and higher proliferative capacity 
. Activation of terminally-differentiated high avidity T cells is thought to lead to weaker responses due to their limited proliferation. However, high avidity T cells, including central memory cells, are often exhausted due to chronic tumor antigen stimulation and may exist in a “corrupted” or tolerant state 
. Thus, additional studies are needed to further clarify the role that the T cell activation state plays in tumor clearance in tolerant hosts.
Many studies have also focused on antigen-targeted low avidity T cell activation and function due to the predominance of low avidity T cells in cancer-bearing hosts. Previous studies suggested that low avidity T cells remain ignorant of antigen expression and therefore do not mount a successful tumor-specific response 
, while others suggest that CD4+
T cell help enhances low avidity T cell function allowing for some degree of tumor destruction 
. While those studies attempted to address the role of low avidity T cells in tumor-bearing mice, none had available a source of naïve low avidity CD8+
TCR transgenic T cells specific for a natural tumor antigen. Here, we show that neu-specific low avidity T cells are capable of activation when adoptively transferred into a non-tolerant environment. In neu
-N mice, however, low avidity T cells were unable to reach a level of activation necessary for adequate IFNγ secretion, integrin upregulation, or tumor infiltration. As a result, these low avidity cells did not affect tumor clearance, and their activation and migration status was not altered by Treg reduction. The difference in low avidity T cell function in the non-tolerant versus tolerant environment is likely due in part to the presence of effective CD4+
T cell help in non-tolerant mice. Our finding agrees with other studies that show that with sufficient CD4+
T cell help, low avidity CD8+
T cells can become active and infiltrate the tumor microenvironment 
. Since low avidity T cells were able to migrate into the tumor, but still were not as potent as the high avidity T cells in rejecting tumor in non-tolerant FVB/N mice, it is likely that alternative mechanisms are blocking optimal low avidity T cell activation even in a more immune permissive setting.
Treg subsets effectively suppress high avidity T cells in vitro
, yet we only observed a trend in suppression with an in vivo
suppression assay. One possibility that explains why the in vivo
suppression assay results failed to reach significance is the technical difficulty of isolating sufficient numbers of Tregs that maintain stable function after adoptive transfer. It is also possible that systemic evaluation of CD25low
Treg suppression of high avidity T cells is not an optimal assay since CD25low
Tregs likely function best within the tumor micro-environment, where they preferentially migrate 
. Some studies have shown that tumor-specific T cells are tolerized at the tumor site. Although technically difficult, future in vivo
studies will attempt to address where CD4+
Tregs affect tolerance.
Emerging data demonstrate that even among CD4+
Tregs, there are sub-populations with different functions and sites of activity. To date, there are only a few studies that have attempted to characterize an inflammation-seeking and highly suppressive CD25low
Treg effector/effector memory population 
. The surface markers expressed by these Tregs point towards their critical suppressive role in tumor-bearing hosts. High β1 integrin, LFA-1, and CXCR3 expression on CD25low
Tregs allow for increased blood vessel extravasation and migration into the tumor, while the low levels of CD62L indicate a migratory phenotype. The importance of ICOS expression by Tregs has been well documented and may be responsible for increased IL-10 production 
, while CD44, CTLA-4, and GITR are also expressed highly on antigen experienced or activated Tregs 
. Although a few studies have shown that this CD25low
suppressor T cell population also express CD103 
, we were unable to detect any Treg CD103 expression on CD25low
Tregs in neu
-N mice. Other groups have describe Tbet expression by CD25low
Treg cells 
. The CD25low
Tregs in our model also express high levels of CXCR3 and Tbet. (Data not shown) Taken together, these findings suggest that the CD25low
Tregs identified in this study likely represent an effector/memory Treg subpopulation similar to that described by other groups.
Treg depletion is believed to be an important strategy in enhancing cancer vaccine efficacy in tumor-bearing hosts. Many pre-clinical and clinical studies have successfully used a variety of Treg-depleting agents to reduce tolerance and activate CD8+
T cell populations in the setting of tumor development 
. Treg depletion strategies strive to deplete all Treg populations because we do not understand which populations predominantly prevent T cell activation in tumor-bearing hosts. Therefore, the study of Treg subpopulations may allow the development of new Treg depleting agents that remove the Tregs responsible for inhibiting cancer antigen-targeted T cells and preserve Treg populations required for inhibiting autoimmune disease. Thus, the surface activation markers expressed on the subset of Tregs that migrate to the tumor may be important targets for future Treg depletion strategies to improve vaccine therapy. This study also verifies the importance of this CD25low
Treg population in a clinically relevant model of breast cancer.
In addition to its CD25low
Treg-depletion effects, Cy may also have other beneficial effects in a tumor-bearing host. A few studies have shown that in certain models, Cy increases type I IFN production and activates DCs that increase effector cell populations 
. These findings may explain the apparent longer-term effects of Cy treatment on the persistence and activation of adoptively transferred high avidity T cells into FVB/N mice when compared with the short-term effects on the same T cells adoptively transferred into neu
-N mice. It is possible that Cy depletes tolerogenic DCs or other antigen presenting cells responsible for activating Tregs 
. The resulting shift in DC homeostasis would ultimately lead to increased activation of tumor-specific CD8+
T cells. This effect would likely be short-term in neu
-N mice due to the regeneration of such antigen presenting cell populations, whereas the effect would be longer-term in non-tolerant FVB/N mice. In addition, by increasing type I IFNs and DC activation, Cy may increase T cell survival and memory development 
. Recently there has been a link postulated between lymphopenia and autoimmune disease. Since Cy does cause mild lymphopenia, in the neu
-N system, a mild form of autoimmunity may be induced leading to tumor destruction 
. Finally, a recent study suggested that Cy may promote the differentiation of non-Treg CD4+
T cells into pro-inflammatory T helper 17 cells that enhance anti-tumor immunity 
. Therefore, Cy may have multiple tolerance-reducing effects that lead to enhanced anti-tumor activity.
In conclusion, these findings demonstrate for the first time that high avidity antigen-targeted CD8+ T cells are regulated predominantly by Tregs in cancer bearing hosts. In addition, a subset of effector/memory Tregs are largely responsible for suppressing high avidity T cells in the tumor microenvironment. Low avidity tumor-specific T cells can function in a non-tolerant host following antigen-targeted vaccination but require as of yet undefined activation signals to achieve adequate function in tolerant hosts and optimal function in non-tolerant hosts. Thus, through the use of high avidity and low avidity T cell populations specific for the same antigen, the specific requirements for T cell activation, trafficking, and function within the tumor micro-environment can be delineated under different tumor-bearing conditions. Ultimately, this information will lead to improved combinatorial immune based approaches for cancer therapy.