Frontline: An in-depth evaluation of the production of IL-2 by antigen-specific CD8 T cells in vivo

IL-2 is an important cytokine that is capable of inducing both proliferation and apoptosis of activated T cells. CD4 T cells are thought to be the major producers of IL-2, but CD8 T cells also produce copious amounts of this cytokine. However, our current understanding regarding the kinetics of IL-2 production by antigen-specific CD8 T cells, and the proportion of these cells that produce IL-2 in vivo, is extremely limited. We now demonstrate that virus-specific CD8 T cells initiate IL-2 production by 6 h post-infection and prior to cell division in vivo. Interestingly, peak levels of IL-2 production were achieved very early during the response and prior to the proliferative peak. We also show -- using transgenic mice expressing herpes simplex virus-1 thymidine kinase under the control of the IL-2 promoter -- that, unlike what has been reported for antigen-specific CD4 T cells, the majority of antigen-specific CD8 T cells produce IL-2 during primary as well as secondary immune responses in vivo.     

Visualizing the course of antigen-specific CD8 and CD4 T cell responses to a growing tumor

Spontaneous tumors frequently express antigens that can be recognized by the immune system but nevertheless manage to evade immune surveillance. To better understand the mechanism of evasion, we followed CD8 and CD4 T cells reacting against a subcutaneously growing tumor, modified to express influenza hemagglutinin (HA) as surrogate tumor antigen. Adoptive transfer of 8,000 antigen-specific CD8 T cells was sufficient to protect against challenge with 1x10(6) tumor cells, while larger numbers of T cells rejected established tumors. HA-specific CD4 T cells could not reject tumors on their own but helped rejection by CD8 T cells. Rejection of the tumor coincided with prolonged survival of expanded antigen-specific CD8 and CD4 T cells, while a failing anti-tumor response was accompanied by transient expansion followed by rapid elimination of antigen-specific T cells. Thus, a highly immunogenic tumor can evade immune surveillance because of an insufficient number of tumor-specific T cells and antigen overload, resulting in exhaustion of the immune response. In this scenario, adoptive immunotherapy rather than vaccination promises successful treatment.     

Cooperation between CD4 and CD8 T cells for anti-tumor activity is enhanced by OX40 signals

The relative contribution of OX40 (CD134) to priming of CD8 T cells in complex systems where CD4 and CD8 cells respond and cooperate together is not clear. We previously found that OX40 expressed on tumor-reactive CD8 T cells controls their initial persistence when adoptively transferred in vivo and is required for delayed tumor growth. We now show that exogenous stimulation of OX40 with agonist antibody augments its ability to suppress the growth of new as well as established tumors, correlating with marked expansion of adoptively transferred CD8 T cells. Concomitantly, anti-OX40 strongly enhanced the number of tumor antigen-reactive CD4 T cells. Moreover, the augmented accumulation of CD8 T cells was prevented in animals lacking MHC class II or depleted of CD4 cells and did not occur in OX40-deficient animals receiving wild-type CD8 cells, demonstrating that non-CD8 cells are the major target of OX40 signals. These results suggest that while OX40 signaling to a CD8 T cell can control its expansion, OX40 expressed on non-CD8 cells strongly influences CD8 priming and in vivo activity. OX40 therefore represents an important signal for allowing effective cooperation between CD4 and CD8 cells and for promoting cell interplay and tumor rejection where CD8 activity is limiting.     

The role of TCR stimulation and TGF-beta in controlling the expression of CD94/NKG2A receptors on CD8 T cells

Following antigen recognition, murine CD8 T cells express CD94/NKG2A receptors. Our results show that this up-regulation occurs rapidly in vitro and is accompanied by an approximately 8-fold increase in CD94 and approximately 125-fold increase in NKG2A mRNA. In contrast, only a twofold increase in NKG2C mRNA is noted. The addition of TGF-beta, but not IL-10, IL-12 or IL-15, leads to a further increase in cell membrane expression of these receptors, as well as a approximately 6-fold increase in mRNA for both chains. TGF-beta also increases CD94/NKG2A expression on memory CD8 T cells that are re-exposed to antigen. The effect of TGF-beta on increasing CD94/NKG2A expression on both naive and memory CD8 T cells occurs only when there is a concurrent stimulation through the TCR. In contrast, TGF-beta does not increase expression of CD94/NKG2A on resting or activated NK cells. We also show by using purified CD8 T cells, that TGF-beta acts directly on these cells. These results implicate a role for both antigen and TGF-beta in increasing expression of inhibitory CD94/NKG2A receptors on CD8 T cells.     

Effector CD8+ T cells activated in vitro confer immediate and long-term tumor protection in vivo

We studied the ability of CD8+ T cells activated in vitro to mediate tumor protection after transfer into adoptive hosts. TCR transgenic CD8+ T cells were activated in culture with DC and specific peptide antigen, and briefly expanded in IL-2 containing medium. Cultured cells acquired a CD44hiCD62Llo phenotype, and following in vivo transfer they preferentially homed to non-lymphoid tissues and spleen. In vivo, their numbers declined between day 0 and day 20, and then remained relatively stable from day 20 to day 90. Over time, many of the injected cells re-expressed CD62L, and acquired the ability to localize to secondary lymphoid organs. Transferred T cells underwent low-level proliferation, expressed IL-7Ralpha and IL-15Rbeta, were cytotoxic in vivo, and retained the ability to produce IL-2, IFN-gamma and TNF-alpha upon ex vivo restimulation. In addition, transferred T cells conferred a high degree of tumor protection, which was greatest immediately after injection, and remained significant even when tumor was given 90 days after T-cell transfer. We conclude that in vitro generated effector T cells can mediate immediate and long-term tumor protection, and develop into long-lived memory T-cell populations in vivo.     

A novel peripheral CD4+CD8+ T cell population: Inheritance of CD8α expression on CD4+ T cells

In this study we show the inheritance of a CD4⁺CD8⁺ peripheral T cell population in the H.B15 chicken strain. A large proportion of αβ T cells in peripheral blood (20–40%), spleen (10–20%) and intestinal epithelium (5–10%) co-express CD4 and CD8α, but not CD8β. CD4⁺ CD8αα cells are functionally normal T cells, since they proliferate in response to mitogens and signals delivered via the αβT cell receptor as well as via the CD28 co-receptor. These cells induce in vivo a graft versus host-reaction, providing further evidence for their function as CD4⁺ T cells. The CD4⁺CD8αα T cell population was found in 75% of the first progeny and in 100% of further progenies, demonstrating that co-expression of CD4 and CD8 on peripheral T cells is an inherited phenomenon. In addition, cross-breeding data suggest a dominant Mendelian form of inheritance. The hereditary expression of CD8α on peripheral CD4⁺ T cells in chicken provides a unique model in which to study the regulation of CD4 and CD8 expression.     

The degree of CD8 dependence of cytolytic T cell precursors is determined by the nature of the T cell receptor (TCR) and influences negative selection in TCR-transgenic mice

Although much has been learned about CD8 structure-function properties, it has so far not been tested whether the nature of the TCR is sufficient to transfer the property of CD8 dependence versus non-dependence to CD8⁺ cytotoxic T lymphocytes (CTL) and their precursors differentiating in T cell receptor (TCR)-transgenic (Tg) mice. In the present study, we compared the characteristics of dependence on CD8 for stimulation of CTL precursors and antigen-specific cytolysis by CD8⁺ T cells from two TCR-Tg mice expressing respectively the TCR (Tg) from a “CD8-dependent” and from a “CD8-independent” CTL clone, which were both reactive against the H-2K^b alloantigen and originated from H-2^k mice. The results indicate that the property of the Tg⁺CD8⁺ cells from H-2^k TCR-Tg mice corresponds to that of the CTL clone of origin, demonstrating that it is linked to the nature of the TCR. Consistent with this property, Tg⁺CD4⁺ cells could also differentiate into H-2K^b-specific CTL when originating from the “CD8-independent”, but not from the “CD8-dependent” Tg-TCR. The influence of the property of “CD8 dependence” on negative selection occurring in TCR-Tg H-2^klb mice was apparent at two levels: (i) in the thymus, the extent of deletion was much more pronounced for the “CD8-independent” TCR-Tg mice; (ii) in the periphery, Tg^+(hi) cells with low to negative CD8 expression were present for the “CD8-dependent” Tg-TCR, whereas only Tg⁺CD4^?CD8^? cells with low surface Tg-TCR and CD3 expression were found for the “CD8-independent” Tg-TCR, indicating that Tg⁺CD4^?CD8^? cells are susceptible to tolerance induction involving TCR/CD3 surface down-modulation. Furthermore, different in vitro conditions led to H-2K^b-induced stimulation of Tg⁺CD4^?CD8^? cells to differentiate into CTL detected in an anti-TCR clonotypic monoclonal antibody redirected cytolysis assay. Culture in interleukin-2 of H-2^klb Tg⁺CD4^?CD8^? cells was sufficient to induce CTL activity in the “CD8-independent” model, whereas stimulation with cells which overexpressed H-2K^b was required in addition to interleukin-2 to induce CTL differentiation in the “CD8-dependent” model. These data suggest that peripheral Tg⁺CD4^?CD8^? cells present in a situation of in vivo tolerance to H-2K^b can still be triggered by H-2K^b with a sensitivity correlated with the degree of CD8 dependence.     

Defective lymphokine production by most CD8+ and CD4+ tumor-specific T cell clones derived from human melanoma-infiltrating lymphocytes in response to autologous tumor cells in vitro

Human melanomas are infiltrated by tumor-reactive T lymphocytes. However, the ability of these cells to elicit a specific anti-tumor response in vivo remains to be established. Because lymphokine production is critical for T cell functions, we have analyzed the capacity of melanoma-specific tumor-infiltrating lymphocyte (TIL) clones to produce major lymphokines: interleukin-2 (IL-2), interferon-γ (IFN-γ) and interleukin-4 (IL-4), as well as tumor necrosis factor (TNF), in response to direct antigen presentation by autologous and allogeneic tumor cells. We report here that, upon stimulation by autologous melanoma cells, all TIL clones secreted TNF but only a few of them produced significant amounts of IL-2, IL-4 or IFN-γ. Nonetheless, all these clones consistently produced two or three of these last lymphokines upon stimulation with phorbol myristate acetate and calcium ionophore, as well as IL-2 upon CD3 stimulation, showing the existence of three lymphokine profiles among them: Th1, Th0 and a profile characterized by IL-2 and IL-4, but not IFN-γ secretion. Stimulation of TIL clones by allogeneic melanoma lines sharing the appropriate HLA-peptide complexes revealed that defective IL-2 production seemed to be a constant feature for some clones, while it was, for other clones, dependent on the antigen-presenting tumor cells. For this last type of clone, we further showed that defective IL-2 induction resulted from an LFA-3 defect of some melanoma cells or from distinct yet undefined defects of other melanoma lines. Our data suggest that defective lymphokine secretion may be an essential component of the in vivo failure of melanoma-reactive TIL to control tumor development. Interestingly both CD4⁺ and CD8⁺ TIL clones from one patient were fully activated by the autologous melanoma cells in vitro, supporting a potential role of such TIL in spontaneous or induced tumor rejection.     

Maturation of CD4-8- alpha/beta TCR+ T cells induced by CD2-stimulation in vivo and in vitro.

Newly generated ('virgin') rat thymocytes of the immature CD4+8+ double positive (DP) subset were treated in suspension culture for 2 days with the stimulatory pair of anti-CD2 monoclonal antibodies OX-54 and OX-55. Approximately 50% of the recovered cells had downregulated CD4 and CD8 and upregulated the T cell antigen receptor (TCR). CD2-stimulated, but not control thymocytes proliferated in response to TCR plus IL-2 stimulation. In vivo, postnatal injection of OX-54/55 led to a dramatic and selective increase in functionally mature CD4-CD8- double negative (DN) alpha/beta--TCR(high) thymocytes and peripheral T cells. These findings show that CD2 stimulation can promote T cell differentiation and suggest that DN TCR(high) thymocytes can be generated from DP thymocytes via alternative pathways of T cell maturation.     

Persistent expression of CD94/NKG2 receptors by virus-specific CD8 T cells is initiated by TCR-mediated signals

Subsets of CD8 T cells express receptors that are critical in regulating the activity of NK cells. To characterize the expression of these receptors on CD8 T cells we made use of transgenic mice that express a H-2Kb restricted TCR specific for the immunodominant epitope located within the HSV-1 glycoprotein B (gB). Few naive gB-specific T cells express Ly49 or CD94/NKG2 receptors. Following acute infection of C57BL/6 mice with either HSV-1 or a recombinant influenza virus that encodes the gB determinant, gB-specific T cells showed a dramatic upregulation of CD94/NKG2 receptors. Moreover, gB-specific CD8 T cells that expressed CD94/NKG2 receptors were also found to express another NK receptor, KLRG1. We established that while Ag-stimulated gB-specific CD8 T cells primarily express inhibitory isoforms of CD94/NKG2 receptors, these cells remain capable of producing gammaIFN upon peptide stimulation. While peak CD94/NKG2 expression on gB-specific cells was reached 2-3 days following infection, it remained elevated beyond 60 days post-infection with either HSV-1 or a gB-expressing recombinant influenza virus. The data imply that the prolonged expression was not due to persistence of replicating virus and suggest that while recognition of the cognate Ag is necessary to trigger expression of CD94/NKG2 receptors, it is not required for their continued expression on memory T cells.     

Augmentation by transferrin of IL-2-inducible killer activity and perforin production of human CD8+ T cells.

The effects of human transferrin (Tf) on lymphokine (IL-2)-activated killer (LAK) induction from blood lymphocytes of healthy donors was examined. LAK cells were induced by 6-day incubation in medium with recombinant human IL-2 of lymphocytes, and their cytotoxic activity was assessed by measuring 51Cr release from NK-resistant Daudi cells. Tf alone did not induce any LAK activity, but in combination with IL-2, it augmented LAK induction dose- and time-dependently. This augmenting effect was completely abolished by pretreatment with anti-Tf antiserum. Tf augmented the proliferative response of lymphocytes to IL-2 and their expressions of receptors for IL-2 and Tf. CD8+ T cells were isolated from purified blood lymphocytes using antibody-bound magnetic beads. Addition of Tf to cultures of CD8+ cells resulted in significant augmentation of killer cell induction and perforin (PFP) production after 4 days stimulation with IL-2. These results indicate that Tf is important in generation of IL-2-inducible killer properties and PFP activity of human CD8+ T cells.     

TRAF2 regulates T cell immunity by maintaining a Tpl2-ERK survival signaling axis in effector and memory CD8 T cells

Generation and maintenance of antigen-specific effector and memory T cells are central events in immune responses against infections. We show that TNF receptor-associated factor 2 (TRAF2) maintains a survival signaling axis in effector and memory CD8 T cells required for immune responses against infections. This signaling axis involves activation of Tpl2 and its downstream kinase ERK by NF-κB-inducing kinase (NIK) and degradation of the proapoptotic factor Bim. NIK mediates Tpl2 activation by stimulating the phosphorylation and degradation of the Tpl2 inhibitor p105. Interestingly, while NIK is required for Tpl2-ERK signaling under normal conditions, uncontrolled NIK activation due to loss of its negative regulator, TRAF2, causes constitutive degradation of p105 and Tpl2, leading to severe defects in ERK activation and effector/memory CD8 T cell survival. Thus, TRAF2 controls a previously unappreciated signaling axis mediating effector/memory CD8 T cell survival and protective immunity.     

Phenotypic and functional analysis of LCMV gp33-41-specific CD8 T cells elicited by multiple peptide immunization in mice revealed the up-regulation of PD-1 expression on antigen-specific CD8 T cells

The phenotype and function of antigen-specific CD8 T cells are closely associated with the efficacy of a therapeutic vaccination. Here we showed that multiple immunizations with LCMV gp33-41 peptide （KAV） in Freund＇s adjuvant could induce KAV-specific CD8 T cells with low expression of CD127 and CD62L molecules. The inhibitory receptor PD-1 was also expressed on a substantial part of KAV-specific CD8 T cells, and its expression level on KAV-specific CD8 T cells in spleen and lymph nodes was much higher when compared to those in peripheral blood. Furthermore, KAV-specific CD8 T cells could specifically kill KAV-pulsed target cells in vivo but the efficiency was low. These data suggest that prime-boost vaccination schedule with peptide in Freund＇s adjuvant can elicit antigen-specific CD8 T cells of effector-like phenotype with partial functional exhaustion, which may only provide short-term protection against the pathogen. Cellular ＆ Molecular Immunology.     

The inhibitory leukocyte-associated Ig-like receptor-1 (LAIR-1) is expressed at high levels by human naive T cells and inhibits TCR mediated activation

Human leukocyte-associated Ig-like receptor-1 (LAIR-1) is a transmembrane glycoprotein with a single extracellular Ig-like domain and a cytoplasmic tail containing two immunoreceptor tyrosine-based inhibition motifs (ITIMs). It is constitutively expressed on the majority of human mononuclear leukocytes and functions as an inhibitory receptor. In this study, we show that freshly isolated peripheral blood T cells are heterogeneous in their expression levels of LAIR-1. We have found that naive T cells express the highest levels of LAIR-1, even more than memory cells. The cross-linking of LAIR-1 inhibits T cell receptor (TCR) mediated signals in freshly isolated human naive T cells and whole populations of CD4⁺ or CD8⁺ T cells. TCR cross-linking increased cell surface expression of LAIR-1 in a process that requires p38 MAP kinase and ERK signaling. Altogether, these results indicate that LAIR-1 is capable of negatively regulating T cell functions, and its high level of expression by naive T cells suggests that it may function at an early stage in the development of an immune response.     

Enhancement of proliferation and downregulation of TRAIL expression on CD8+ T cells by IL-21

Mounting evidence indicates that the cytokine IL-21 can significantly enhance the survival of CD8(+) T lymphocytes. Given that CD4(+) T lymphocytes constitute the main cellular source for IL-21 in vivo, it is tempting to speculate a direct role in mediating the "help" provided by these CD4(+) T cells to the CD8 response. A new report in this issue of the European Journal of Immunology advances this notion by showing that CD8(+) T cells lacking the IL-21 receptor phenocopy those primed in the absence of CD4(+) T cells (the so-called "helpless" CD8(+) T cells) in their induction of the pro-apoptotic factor TRAIL. This finding helps to define the role of IL-21 in the CD8 response, and raises new questions relevant for achieving a broader understanding of this multifunctional cytokine.     

TGF‐β downregulates KLRG1 expression in mouse and human CD8+ T cells

The inhibitory receptor killer cell lectin‐like receptor G1 (KLRG1) and the integrin α_E (CD103) are expressed by CD8⁺ T cells and both are specific for E‐cadherin. However, KLRG1 ligation by E‐cadherin inhibits effector T‐cell function, whereas binding of CD103 to E‐cadherin enhances cell–cell interaction and promotes target cell lysis. Here, we demonstrate that KLRG1 and CD103 expression in CD8⁺ T cells from untreated and virus‐infected mice are mutually exclusive. Inverse correlation of KLRG1 and CD103 expression was also found in human CD8⁺ T cells‐infiltrating hepatocellular carcinomas. As TGF‐β is known to induce CD103 expression in CD8⁺ T cells, we examined whether this cytokine also regulates KLRG1 expression. Indeed, our data further reveal that TGF‐β signaling in mouse as well as in human CD8⁺ T cells downregulates KLRG1 expression. This finding provides a rationale for the reciprocal expression of KLRG1 and CD103 in different CD8⁺ T‐cell subsets. In addition, it points to the limitation of KLRG1 as a marker for terminally differentiated CD8⁺ T cells if lymphocytes from tissues expressing high levels of TGF‐β are analyzed.     

CD4-negative cytotoxic T cells with a T cell receptor α/βintermediate expression in CD8-deficient mice

Targeted disruption of the CD8 gene results in a profound block in cytotoxic T cell (CTL) development. Since CTL are major histocompatibility complex (MHC) class I restricted, we addressed the question of whether CD8–/– mice can reject MHC class I-disparate allografts. Studies have previously shown that skin allografts are rejected exclusively by T cells. We therefore used the skin allograft model to answer our question and grafted CD8–/– mice with skins from allogeneic mice deficient in MHC class II or in MHC class I (MHC-I or MHC-II-disparate, respectively). CD8–/– mice rejected MHC-I-disparate skin rapidly even if they were depleted of CD4⁺ cells in vivo (and were thus deficient in CD4⁺ and CD8⁺ T cells). By contrast, CD8+/+ controls depleted of CD4⁺ and CD8⁺ T cells in vivo accepted the MHC-I-disparate skin. Following MHC-I, but not MHC-II stimulation, allograft-specific cytotoxic activity was detected in CD8–/– mice even after CD4 depletion. A population expanded in both the lymph nodes and the thymus of grafted CD8–/– animals which displayed a CD4^?8^?3^intermediateTCRα/β^intermediate phenotype. Indeed its T cell receptor (TCR) density was lower than that of CD4⁺ cells in CD8–/– mice or of CD8⁺ cells in CD8+/+ mice. Our data suggest that this CD4^?8^?T cell population is responsible for the CTL function we have observed. Therefore, MHC class I-restricted CTL can be generated in CD8–/– mice following priming with MHC class I antigens in vivo. The data also suggest that CD8 is needed to up-regulate TCR density during thymic maturation. Thus, although CD8 plays a major role in the generation of CTL, it is not absolutely required.