| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print April 23, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Molecular and Cellular Biology, Research Institute, Sunnybrook and Women’s College Health Sciences Center, Toronto, Canada; Toronto-Sunnybrook Regional Cancer Center, Canada; and Departments of Medicine and Medical Biophysics, University of Toronto, Canada
1Correspondence: Division of Molecular and Cellular Biology, Research Institute, S-116A, Research Building, Sunnybrook and Women’s College Health Sciences Center, 2075 Bayview Avenue, Toronto, Ontario, Canada M4N 3M5. E-mail: spanerd{at}srcl.sunnybrook.utoronto.ca
 |
ABSTRACT |
|---|
2b provide a "proof of principle" in human patients, that agents which counter the gene programs of cancer cells, causing them to intrinsically resist tumor-reactive T cells, may improve significantly the efficacy of cancer vaccines.
Key Words: human • tumor immunity • cytokines • signal transduction
|
INTRODUCTION |
|---|
|
|
REQUIREMENTS FOR T CELL-MEDIATED CONTROL OF TUMOR CELLS |
|---|
are important for the development of type 1 (or TH1/TC1) immunity, which is believed to be required for therapeutic anti-tumor effects [2
]. Fully stimulated CTLs proliferate and differentiate into effector cells that make TH1 cytokines, such as IFN-
, and kill tumor cells. An effective cancer vaccine is expected to provide tumor antigens and costimulatory signals in a manner that can activate tumor-reactive T cells to levels that approach those of successful anti-viral TH1/TC1 responses (of the order of 1–10/100 CD8+ T cells [8
]) for sufficient time (probably of the order of several months [9
]) to achieve clinically meaningful results.
Tumor antigens
Most tumors express antigens that are recognizable by CTLs [10
]. Using a number of experimental approaches, several groups have identified molecular targets in different cancers [11
]. These antigens may be classified into groups that include cancer-testis antigens, such as the melanoma-associated antigens [12
], which are encoded by genes expressed in fetal tissues and solid tumors but not in normal tissues except for testis; antigens arising from genes that are mutated during oncogenesis and encode proteins that produce novel peptides upon proteasomal degradation, such as mutated mum-1 in melanoma [8
]; and differentiation antigens that are expressed only in the cell of origin of the cancer, such as gp100 antigens in melanoma cells that are also expressed by normal melanocytes [13
]. For cancers in which defined antigens have not yet been identified, vaccines can still be made directly from the patients’ own tumor cells that express multiple (albeit unknown) antigens [14
].
Tumor-reactive CTLs
CTLs that recognize tumor antigens and can respond potentially to vaccines are present in many patients. Numbers of these cells are sometimes quite low as a result of tolerance mechanisms (as many tumor antigens are also self-antigens), previous treatment with cytotoxic chemotherapy, and factors produced by tumor cells during disease progression [15
]. In some patients, tetramers of HLA molecules folded around a defined epitope can be used to enumerate tumor-reactive T cells by flow cytometry [16
]. Using such assays, others and I [13
] have shown that T cells reactive to gp100 and other melanoma antigens exist in some HLA-A*0201+ healthy controls, melanoma patients, and patients with vitiligo (a disease caused by autoimmunity against normal melanocytes). Moreover, these T cells can be increased in number by melanoma antigen-containing vaccines (Fig. 1A)
.
In the absence of defined antigens or suitable HLA alleles, enzyme-linked immunospot (ELISPOT) assays can be used to enumerate tumor-reactive T cells by their localized production of cytokines after stimulation by tumor cells [17
]. Using such cellular ELISPOT assays, co-workers and I [4
] have demonstrated the presence of tumor-reactive T cells in 
60% of CLL patients and increases in CLL-reactive T cell frequencies after autologous tumor cell vaccines.
Cancer patients still often fail to develop meaningful, therapeutic responses, even using vaccines composed of defined antigens presented by powerful antigen-delivery systems that cause tumor-reactive T cells to comprise the majority of circulating CD8+ T cells [18 ]. Several aspects of cancer cell biology may account for this seemingly paradoxical situation. For example, optimal infiltration of solid tumors and effector function of CTLs appear to require CD80 and CD86 expression by tumor cells [19 ]. Unfortunately, the expression of costimulatory molecules by tumor cells is often very low [20 ]. Importantly, cancer cells also produce commonly a number of immunosuppressive factors that inhibit the effects of tumor-reactive T cells and make tumor cells resistant to the lytic effects of CTLs.
|
IMMUNOSUPPRESSIVE FACTORS IN CANCER PATIENTS |
|---|
TGF-ß
The TGF-ß family is composed of three ubiquitously expressed genes [25
]. TGF-ß signals through a complex of type I and II transmembrane serine/threonine receptor kinases (TGF-ßRI/RII) found on most cells. Members of the SMAD protein family (the name is derived from the Sma and MAD gene homologues in Caenorhabditis elegans and Drosophila melanogaster) [26
] mediate signal transduction through TGF-ßRs. SMAD2 and SMAD3 are phosphorylated by activated TGF-ßRI and form a heterotrimeric transcription factor with SMAD4. SMAD6 and SMAD7 prevent receptor-mediated phosphorylation of SMAD2 and SMAD3 and interfere with nuclear translocation of the transcription factor. Many genes are transcriptionally regulated by TGF-ß. In particular, TGF-ß is a potent inducer of genes that inhibit cell-cycle progression, such as p21CIP1 [25
].
In the early stages of cancer, TGF-ß behaves as a tumor suppressor gene, because of its inhibitory effects on the cell cycle. Alterations of TGF-ßR signaling components accompany disease progression and presumably allow tumor cells to escape the inhibitory effects of TGF-ß. TGF-ßR expression is decreased in some tumors, such as CLL [27 ], and inactivating TGF-ßRII gene mutations are found in gastric cancers, gliomas, and 20–25% of all colon cancers [28 ]. Mutations in SMAD2 and SMAD4 genes are detected in several carcinomas, especially the pancreas [29 ].
Circulating TGF-ß levels are often increased in advanced-stage cancer patients [30 ]. TGF-ß2 mRNA is expressed in all metastatic and 94% of deep primary melanoma lesions but not in benign nevi [31 ]. In colorectal cancer, plasma TGF-ß1 levels correlate with extent of disease and decrease after curative surgical procedures [32 ]. In lung [33 ], prostate [34 ], gastric [35 ], and breast cancers [36 ], elevated TGF-ß levels predict a poor outcome. Taken together, these observations indicate that tumor cells often produce TGF-ß when they no longer respond to its inhibitory effects. However, normal cells in the tumor microenvironment remain susceptible to the effects of TGF-ß, and this is especially true for cells of the immune system.
TGF-ß suppresses proliferative T cell responses to IL-2 and mitogens [37
] and expression of perforin, an important mediator of cellular cytotoxicity by CTLs and natural killer (NK) cells [38
]. TGF-ß also stimulates macrophage arginase activity [39
], causing reduced extracellular L-arginine levels, decreased expression of the CD3
chain by T cells and impaired TCR-mediated responses [40
].
VEGF
The VEGF protein family consists of four isoforms that signal through Flt, Flk-1/KDR, and Flt-4 [41
]. VEGF mRNA and protein are made by many hematological and solid tumors [42
], and elevated plasma VEGF levels tend to correlate with aggressive disease.
VEGF appears not to suppress human T cells directly [43 ]. However, VEGF inhibits the development of mature DCs that express high levels of costimulatory molecules and elicit strong primary and secondary T cell responses. Increased numbers of immature DCs are seen in acute and chronic leukemia patients [44 ], conceivably because of VEGF produced by leukemic cells. As tolerance can result from T cell interactions with immature DCs [45 ], and mature DCs sustain the responses of vaccine-activated T cells [46 ], the effects of VEGF on DC development could have a major negative impact on the clinical outcome of a cancer vaccine.
IL-6
IL-6 is produced by a number of primary and transformed cells [47
]. Dimerization of IL-6 [bound to membrane or soluble forms of the chain of the IL-6 receptor (IL-6R)] with the signal-transducing element, gp130/IL-6Rß, leads to activation of nuclear factor of activated T cells and phosphorylation of the Janus tyrosine kinase (JAK) family members, JAK1, JAK2, and TYK2. These tyrosine kinases then phosphorylate the members of the signal transducers and activators of transcription (STAT) family, STAT3, and (to a much lesser extent) STAT1. Homodimers of phosphorylated STAT3 translocate to the nucleus and regulate the transcriptional activity of VEGF and a variety of genes involved in inflammation and apoptosis [48
]. High IL-6 plasma levels correlate with advanced stage and poor prognosis in CLL [49
], other lymphomas [50
], and renal cell cancer (RCC) and ovarian and prostate cancers [51
].
IL-6 inhibits anti-tumor T cell responses by affecting T cell differentiation and APC function. IL-6 causes naïve CD4+ T cells to make IL-4 and become TH2/TC2 cells and inhibits TH1 cell development [52
] by up-regulating expression of the suppressor of cytokine signaling (SOCS) family member, SOCS1, which prevents IFN- signaling [53
]. Activation of STAT3 by IL-6 inhibits the ability of APCs to activate anergic T cells [54
], which has important implications for cancer vaccines, as many tumor-reactive T cells recognize self-antigens and may be in a state of anergy [4
].
PGE2
PGE2 is a potent immunosuppressive factor found in the supernatants of many cultured tumor cells [55
]. It is produced from arachidonic acid and other 20 carbon polyunsaturated fatty acids by cyclooxygenase-2 (COX-2) and prostaglandin E synthase (PGES). COX-2 and PGES expression and activity are increased in a number of hematological and solid tumors and associated often with poor outcome [56
].
Receptor binding of PGE2 activates adenylate cyclase, production of cyclic adenosine monophosphate (cAMP), and the protein kinase A (PKA) pathway [55 ]. cAMP inhibits TH1 cytokine production and increases IL-4 and IL-10 production by T cells. Picomolar concentrations of PGE2 inhibit T cell proliferation [57 ] and (like VEGF) suppress DC function [58 ].
IL-10
IL-10 is made by many primary tumors and cell lines [59
]. Serum IL-10 levels are elevated especially in B lymphomas (likely mirroring the production of IL-10 by activated normal B cells) and can be associated with poor outcome [60
]. The IL-10R is composed of two chains, which bind JAK1 and TYK2, respectively. Like IL-6, STAT3 and small amounts of STAT1 are activated during IL-10 signaling [61
]. IL-10 inhibits TH1 cytokine production and expression of CD80 and CD86 by APCs and tumor cells [62
].
|
REGULATORY CD4+ T (Treg) CELLS AND MACROPHAGES |
|---|
Treg cells that have been activated through their TCRs [65 ] inhibit anti-tumor immunity by cell–cell contact mechanisms involving membrane-bound TGF-ß and cytotoxic T-lymphocyte antigen-4 (CD152) [66 ] or secretion of IL-10, TGF-ß, and IL-4 [67 ]. Markers for Treg cells include CD25, CD152, and the glucocorticoid-induced tumor necrosis factor receptor family-related protein (GITR) as well as the transcription factor, Scurfin, which is the product of the forkhead box P3 gene [65 ]. Increased numbers of Treg cells have been found in the peripheral blood and tumor microenvironment of patients with a variety of solid tumors [68 69 70 71 72 73 ]. The antigens recognized by Treg cells do not seem to differ from cancer antigens recognized by CD4+ effector cells, although only the cancer-testis antigen, LAGE1, has been described as a target for Treg cells in human patients to date [74 ]. The reason for the increased numbers of Treg cells in advanced-stage cancer patients is not clear. If the physiological function of Treg cells is to prevent the development of autoimmunity, then the increased mass of mainly self-antigens on tumor cells could lead to the expansion of Treg cell numbers. However, CD4+ T cells that have been activated in the presence of IL-10 and TGF-ß can become CD25+ regulatory cells [75 ]. Therefore, tumor cells (by their expression of tumor antigens and production of these factors) may cause antigen-activated T cells to differentiate into Treg cells. The production of immunosuppressive factors from Treg cells (and other cells in the tumor microenvironment [76 ]) would be expected to consolidate further the immunosuppressive effects of tumor cells.
|
ARE TUMOR-DERIVED, IMMUNOSUPPRESSIVE FACTORS A RESPONSE TO IMMUNE SURVEILLANCE? |
|---|
|
SIGNALING PATHWAYS IN ONCOGENESIS |
|---|
RAS
The proto-oncogene, ras, controls a phosphorelay system that leads to the phosphorylation and activation of the serine/threonine mitogen-activated protein kinases (MAPKs), extracellular-regulated kinase (ERK)1 and ERK2 [81
]. RAS is a member of a superfamily of small guanosine 5'-triphosphate (GTP)-binding proteins that are post-translationally modified to localize them to the proper subcellular compartment, which is usually the inner plasma membrane. RAS is active when it binds GTP and inactive when it binds guanosine 5'-diphosphate, and as RAS has little intrinsic GTPase activity, RAS activation is controlled by a number of activating guanine nucleotide exchange factors (GEFs) and inactivating GTPase-activating proteins (GAPs). GEFs couple receptors, such as receptor tyrosine kinases (RTKs) to the RAS pathway. GTP-bound RAS binds and activates members of the RAF family of protein serine/threonine kinases. In turn, activated RAF proteins phosphorylate and activate MAPK kinases 1 and 2 (MEK1 and MEK2), which are dual-specificity kinases that phosphorylate ERK1 and ERK2.
The RAS pathway is of central importance to carcinogenesis. Some 20% of human tumors have point mutations in ras genes, which cause the accumulation of active GTP-bound RAS [82 ]. RAF members are also activated frequently by mutations. For example, 66% of melanomas harbor RAF mutations [81 ]. Activated RAS signaling pathways also reflect the constitutive action of RTKs, such as the epidermal growth factor receptor (EGFR) and HER2/neu, in many solid tumors.
ERK substrates include transcription factors, such as activated protein 1, which regulate genes required for transformation and invasion of surrounding tissue [83 ]. Other genes that are activated by RAS effectors prevent apoptosis and make tumor cells more resistant to killing by vaccine-activated T cells. Importantly, a number of the immunosuppressive factors, described above, are made directly as a result of activation of the RAS signaling pathway. Transfection of activated RAS into cell lines results in the production of VEGF [84 ], TGF-ß [85 ], COX-2, and PGES [86 ]. RAS signaling also inhibits TGF-ß effects through ERK-mediated phosphorylation of SMAD2 and SMAD3, which prevents their nuclear translocation [87 ].
PI-3K-AKT
The serine/threonine kinase, AKT, the cellular homologue of the retroviral oncogene, v-akt (also known as PKB), is activated by PI-3K [88
]. Inactive PI-3K, composed of a p85 regulatory subunit constitutively bound to a p110 catalytic subunit, is recruited by an activated RTK to the cell membrane, where p110 becomes active through proximity to its lipid substrates and also conformational changes that relieve inhibition by p85. Activated PI-3K generates the second messenger, PI(3,4,5)-trisphosphate (PIP3) from PI(4,5)-bisphosphate (PIP2). PIP3 recruits AKT to the cell membrane [where it is phosphorylated and activated by 3-phosphoinositol-dependent protein kinase-1 (PDK1) and PDK2] by binding to its pleckstrin-homology domain [89
].
AKT positively regulates cell proliferation, growth, and survival [88 ]. For example, AKT phosphorylates and inhibits the products of proapoptotic genes, such as BAD and members of the FOXO family of forkhead transcription factors. AKT also inhibits glycogen synthase kinase 3 (GSK3), which regulates the degradation of cyclin D1 and MYC, and the FOXO transcription factors also repress cyclin D1 expression and enhance expression of cell-cycle inhibitors such as p27 [90 ]. By activating proteins that drive entry into S phase and repressing cell-cycle inhibitors, AKT promotes cell-cycle progression.
AKT promotes cell growth by phosphorylating and inhibiting tuberin, which binds to hamartin, forming the tuberous sclerosis (TS) complex, an inactiving GAP for the RAS-like small G protein, RHEB, which is an activator of mammalian target of rapamycin (mTOR), which controls cell growth by activating the 70-kDa ribosomal S6 kinase (S6K1) and stimulating protein translation [91 ]. PDK1, which is also activated by PIP3, phosphorylates and activates S6K1 directly.
The major regulator of the PI-3K-AKT pathway is the phosphatase and tensin homologue (PTEN), which is a dual-specificity phosphatase that converts PIP3 back to PIP2 and shuts off PI-3K signaling. PTEN can be inactivated by gene deletion, mutation, transcriptional silencing, or protein instability, which leads to dysregulated activity of AKT and PDK1 [92 ]. PTEN has a mutational frequency second only to the tumor suppressor, p53, in a wide variety of cancers. For example, glioblastomas have a PTEN mutation incidence of 25–60%, which may be related to their inherent resistance to chemotherapy. Mutations that inhibit p85 function [93 ], or amplification of p110 [94 ] or akt genes [95 ], may also activate the PI-3K-AKT pathway during carcinogenesis.
With respect to immunosuppressive factors, PI-3K and AKT increase transcription of VEGF by cancer cells [96 ]. Although its effects on IL-6 and IL-10 expression by tumor cells have not been well studied, the PI-3K-AKT pathway mediates IL-10 production by lipopolysaccharide (LPS)-activated human monocytes and represses production of IL-12, an important mediator of type 1 immunity [97 ]. In addition, the antiapoptotic effects of AKT inhibit killing by CTLs, as well as chemotherapeutic agents [98 ].
STAT3
Constitutive activation of STAT3 (indicated by phosphorylation of specific tyrosine or serine residues) is observed in many human cancers [99
], likely mediated by transforming receptor and cytoplasmic tyrosine kinases (such as EGFR, SRC, or the translocated ets leukemia-JAK2 fusion protein [100
]), or autocrine loops of cytokines, such as IL-6 and IL-10 (described above). Although naturally occurring stat3 gene mutations have not been described, a synthetic mutant can transform mouse fibroblasts [101
]. The pathogenic importance of activated STAT3 is shown by the ability of specific pharmacologic or genetic inhibitors to kill or inhibit the growth of a number of tumor cells in vivo and in vitro [99
]. Activated STAT3 regulates the expression of many genes that promote cell growth and prevent cell death. Regarding immunosuppressive factors, STAT3 causes VEGF production by cancer cells [102
] and controls il10 gene expression in LPS-activated human monocytes [103
], suggesting it may also regulate IL-10 expression by cancer cells. Tumor-derived factors also impair dendritic cell (DC) maturation through the induction of high levels of activated STAT3 [104
].
|
IMMUNOSUPPRESSION AS A CONSEQUENCE OF TRANSFORMATION |
|---|
|
IMMUNOMODULATORS AS CANCER VACCINE ADJUVANTS |
|---|
Targeting the RAS pathway
At the present time, the RAS-RAF-MEK-ERK pathway can be blocked with clinically relevant reagents from the level of activated RTKs down to MEK1 and MEK2. In epithelial or breast cancers, EGFR or HER2/neu, respectively, commonly activate RAS signaling [113
]. Small molecules (such as ZD1839 and OSI-774) or antibodies against the extracellular receptor domains (such as Erbiten or Herceptin) can inhibit these kinases [81
]. Given the role of activated RAS in the production of immunosuppressive factors by tumor cells (described above), these inhibitors could decrease immunosuppressive factor levels and amplify the effects of vaccines directed against antigenic targets on epithelial tumors [114
]. Use of these agents in combination with vaccines has not yet been studied but is appealing, especially as they are unlikely to have negative effects on tumor-reactive T cells.
Direct inhibition of RAS, itself, by antisense oligonucleotides (which prevent its expression) or farnesyltransferase inhibitors (which prevent its localization to the plasma membrane), has been problematic [115 ]. However, inhibitors of downstream effectors of RAS are in clinical trials. For example, BAY 43-9006 can inhibit RAF1 activity [116 ], and CI-1040 inhibits MEK1 and MEK2 [117 ]. These drugs are relatively nontoxic at doses that inhibit ERK activation in PBMCs and could potentially be incorporated into tumor vaccine strategies. Measurement of VEGF, TGF-ß, IL-6, and IL-10 proteins or mRNA in plasma or biopsies, respectively, could be used to gauge the effectiveness of these agents in reducing immunosuppressive factor production by tumor cells. The immunological effects of these drugs have not been well described in humans, but the requirement for the RAS pathway in T cell activitation is a potential concern for using these inhibitors in vaccine strategies [118 ].
Targeting the PI-3K-AKT pathway
Clinically relevant inhibitors of the PI-3K-PDK1-AKT-TS-mTOR pathway are under active development. UCN-01 is a PDK1 inhibitor that can be administered safely to patients [119
], and Wortmannin and Ly294002 are PI-3K inhibitors that have been used extensively in vitro [120
] but not in clinical trials. Rapamycin inhibits mTOR [121
], leading to decreased VEGF production by tumor cells [122
] and increased sensitivity to killing by chemotherapeutic agents (and presumably CTLs) [98
]. Rapamacyin is approved for the treatment of transplant rejection [121
] and might be problematic as a cancer vaccine adjuvant because of its immunosuppressive properties. The same caveat may apply to other inhibitors of the PI-3K pathway, which mediates cell-cycle progression of T cells [121
].
Targeting STAT3
Inhibitors of JAK tyrosine kinases that phosphorylate STAT3 are available. For instance, AG490 is a tyrphostin that inhibits JAK2 especially and has direct anti-tumor effects in vitro and in vivo [123
]. AG490 has synergistic effects with type 1 immunity-promoting agents, such as IL-12 [124
], making its use in combination with cancer vaccines appealing. Cucurbitacin I is a recently identified STAT3 inhibitor that may also prove to have clinical relevance [125
].
Other tyrosine kinases, such as ABL and SRC, phosphorylate STAT3. ABL can be inhibited by imatinib mesylate. This drug has little toxicity but significant activity against cancers with constitutive activation of ABL, such as chronic myelogenous leukemia (CML) and gastrointestinal stromal tumors [126
]. AP23464 is a SRC inhibitor in the early stages of clinical testing [127
]. Through their inhibition of STAT3 phosphorylation, these drugs could potentially sensitize cancers (driven by ABL or SRC) to the effects of vaccine-activated tumor-reactive T cells. The ability of imatinib mesylate to augment the effects of tumor-reactive T cells in cancer patients is supported by the increased T cell production of IFN- in CML patients treated with this drug [128
].
Phosphorylation of Serine 727, in addition to tyrosine phosphorylation, is required for full transcriptional activity of STAT3 [100 ]. Serine kinases of the RAS-RAF-MEK-ERK pathway can potentially phosphorylate STAT3 at this site, which is another reason to consider using RAF and MEK inhibitors (described above) to sensitize tumor cells to killing by vaccine-activated CTLs. For epithelial tumors, antibodies against HER2/neu or EGFR would also inhibit STAT3 activation, as these receptors phosphorylate STAT3 as well as activate RAS [100 ].
Additional, direct inhibitors of activated STAT3 are being developed, such as peptides that block its dimerization, nuclear localization, and ability to mediate gene transcription [129 ]. Agents that promote the activity of natural inhibitors, such as SOCS and protein inhibitors of activated STAT family members [53 ], could potentially also be used in vaccine strategies to inhibit STAT3 in tumor cells.
|
SPECIFIC IMMUNOSUPPRESSIVE FACTOR INHIBITORS |
|---|
TGF-ß
TGF-ß is an appealing target for cancer vaccine strategies, as it is produced by many tumor cells and has profound negative effects on T cells, as exemplified by the generalized autoimmune pathology in TGF-ß knockout mice [130
]. The effects of TGF-ß can potentially be blocked at a number of levels by clinically relevant reagents. Antisense compounds can prevent TGF-ß expression [131
]. Free TGF-ß can be blocked by injections of decorin (a small chondroitin-dermatan sulfate proteoglycan) [132
], soluble ß-glycan, and TGF-ßRII [133
] or humanized monoclonal antibodies against different TGF-ß isoforms [131
]. In preclinical studies, TGF-ß antibodies had synergistic antitumor effects in combination with IL-2 and might have similar effects with vaccines [134
].
VEGF
As VEGF does not appear to inhibit T cells directly [43
], treatments directed against VEGF may improve the outcome of cancer vaccines by improving the function of DCs and maintaining the activated state of tumor-reactive T cells [46
]. The effects of VEGF can be blocked with antibodies [135
] or small molecule VEGF RTK inhibitors [136
]. Indeed, VEGF antibodies have been shown to improve the effectiveness of DC vaccines in tumor-bearing mice [137
] and to reverse DC maturation defects in metastatic lung cancer patients [138
].
IL-6 and IL-10
IL-10 and IL-6 antibodies [47
] could interrupt autocrine loops that lead to STAT3 activation in tumor cells and ineffective type 2 anti-tumor T cell responses [52
]. Drugs such as Rapamycin could also prevent production of these cytokines by tumor cells [121
]. As mentioned, however, Rapamycin’s immunosuppressive effects might be problematic when used in conjunction with vaccines.
PGE2
PGE2 production results from the activation of COX-2 and PGES by cytokines in the cancer microenvironment as well as constitutive RAS activation [139
]. COX-2 expression can be inhibited with nonsteroidal, anti-inflammatory drugs, such as Celecoxib. These agents can inhibit tumor-induced suppression of DC function [58
] and enhance anti-tumor T cell activity after vaccination [140
]. COX-2 inhibition has been shown to be able to prevent immunosuppression in surgical patients [141
] (which is mediated by many of the same factors as in cancer patients [142
]) and would appear to be a rational addition to cancer vaccine strategies [143
], assuming sufficiently low levels of PGE2 are achievable.
|
OPPOSING SIGNALING PATHWAY ACTIVATORS |
|---|
p38 and p38 agonists
The four p38 family members are MAPKs that are activated by the MAPK kinases, MKK3, -4, and -6, which in turn, are activated by members of the MEKK and mixed-lineage protein kinase groups [145
]. Analogous to RAS in the ERK pathway, small G proteins (Rho family members) are the upstream activators of these MKKKs. The MAPK, p38, is activated by chemical and environmental stresses, including hyperosmolarity, chemotherapeutic agents, inflammatory cytokines, and Toll-like receptor (TLR) agonists [146
]. Following activation, p38 regulates the transcription and translation of proinflammatory, antiproliferative, and proapoptotic genes, which counter the effects of STAT3, AKT, and ERK on cell growth and survival. The p38 and ERK pathways can regulate each other by activating a group of MAPK phosphatases (MKPs) [147
]. For example, MKP-1, activated by p38, turns off signaling through RAS by dephosphorylating residues in the activating loops of the ERKs. In general, dominance of the ERK or p38 pathways appears to result from the relative magnitudes of the respective extracellular-activating signals [148
].
Some chemotherapeutic drugs are agonists of the p38 signaling pathway [149
] and can potentially enhance the results of cancer vaccines. As chemotherapeutic agents can be immunosuppressive [150
], drug doses and timing in relation to vaccination are important variables. Chemotherapy administered before vaccination may decrease the production of immunosuppressive factors (through effects on tumor cells and Treg cells) and also create "space" for proliferation of vaccine-activated T cells [151
]. For example, the alkylating agent, cyclophosphamide, has been shown to enhance the subsequent effects of vaccines in clinical trials [152
, 153
]. The microtubule inhibitor and TLR agonist paclitaxel is a vaccine adjuvant in preclinical models [154
]. Its ability to change the immunological profile of breast cancer patients (i.e., to increase IFN- and decrease PGE2 serum levels and increase T cell activity in mixed lymphocyte reactions) [155
] suggests it may prove useful for enhancing vaccine-induced T cell responses.
Other chemotherapeutic agents may be more effective after vaccination. For example, the anthracycline, doxorubicin, enhanced the effects of DC vaccines in neu-transgenic mice bearing tumor implants of neu-expressing breast cancer cells [156 ]. The mechanism of this effect is not clear, but perhaps p38 activation by doxirubicin decreased the production of immunosuppressive factors by the tumor cells and made them more receptive to vaccine-activated CTLs. Moreover, p38 activation in T cells by chemotherapy might protect them from the effects of TGF-ß and IL-10 made by cancer cells in the same way that it protects T cells from these factors during surgical shock [142 ]. For these reasons, low-dose, continuous chemotherapy (along with anti-VEGF antibodies), which is being evaluated as an antiangiogenic strategy [157 ], may also be an effective vaccine adjuvant.
STAT1 and STAT1 agonists
STAT1 is the major transducer of signals from the IFN- and - receptors. In common with STAT3, tyrosine and serine phosphorylation of STAT1 is necessary for it to acquire full transcriptional activity [158
]. STAT1 homodimers bind to -activated sequence elements in the promoters of IFN--responsive genes [159
]. Heterodimers of STAT1 and STAT2 along with p48 (a member of the IFN regulatory factor family of transcription factors) form the trimeric complex IFN-stimulated gene factor 3, which binds to IFN-stimulated response elements in the promoters of IFN--inducible genes. In general, STAT1 has tumor suppressor properties and (in common with p38) promotes the expression of proapoptotic and cell-cycle inhibitory genes and represses cell survival and proliferation genes. STAT1 also mediates the positive effects of IFN on killing by monocytes and lymphocytes, expression of major histocompatibility complex and costimulatory molecules, and antigen presentation [160
] as well as the negative effects of IFN on the expression of genes, such as VEGF [161
]. Thus, the gene program regulated by activated STAT1 is almost the antithesis of activated STAT3, which promotes tumor progression and production of immunosuppressive factors.
Which of the respective gene programs controlled by these two transcription factors becomes dominant is dependent on a number of factors. One is the relative levels of activated STAT1 and STAT3 in a cell, which can be altered by cytokines, such as IFN- or IL-10, respectively. In the presence of high STAT3 levels, STAT1 homodimers and STAT1/STAT2 heterodimers (which transmit the signals from the IFN receptors) may be diluted by STAT1/STAT3 heterodimers, which have a higher association constant [159
]. Analogous to the MKP proteins that regulate the opposing actions of p38 and ERK, SOCS family members regulate opposing JAK-STAT pathways [53
]. For example, STAT3 can suppress STAT1 activation by inducing SOCS1, which competes for binding to JAK kinase domains and prevents tyrosine phosphorylation of STAT1. Diverting the cancer cell gene program to favor STAT1-dominated transcription would be expected to remove a major source of immunosuppressive factors (i.e., those regulated by STAT3) and make the tumor cells more receptive to the lytic effects of CTLs induced by cancer vaccines. Clinically relevant STAT1 agonists that can be used potentially as vaccine adjuvants include Bryostatin, sodium stibogluconate, IL-12, IL-2, and IFN-.
Bryostatin
Bryostatin is a PKC agonist with acceptable clinical toxicity [162
], which can increase T cell cytolytic function [163
] and activate PKC isoforms, including PKC
[164
], which phosphorylates Serine727 and increases transcriptional activity of STAT1 [158
] and also inhibits signaling through STAT3-activating receptors (such as the IL-6R and IL-10R) by phosphorylating serine residues on gp130 [165
]. Bryostatin may thus be a useful vaccine adjuvant (particularly in combination with IFN- to tyrosine-phosphorylate STAT1) through these effects on T cells and PKC. The significant antitumor effects of Bryostatin and a standard chemotherapeutic agent observed in otherwise refractory B cell lymphoma patients [166
], with a time course suggestive of an immune-mediated mechanism of action, suggest that Bryostatin may have clinical relevance as a vaccine adjuvant. Although natural supplies are limited, methods for synthesizing Bryostatin and analogs that stimulate specific PKC isoforms have been developed [167
].
Sodium stibogluconate
The tyrosine phosphatases, Src homology-2-containing tyrosine phosphatase-1 (SHP-1) and SHP-2, shorten the half-life and inhibit the effects of activated tyrosine-phosphorylated STAT1 and STAT3. Blocking the function of these phosphatases would amplify STAT1 and STAT3 activity and depending on the cellular context, could favor STAT1 dominance over STAT3. The SHP-2 inhibitor, sodium stibogluconate [168
], has long been used in the treatment of Leishmania. It has an acceptable toxicity profile, works in part by augmenting cell-mediated immunity [169
] and decreasing IL-5 production by T cells reactive against parasites [170
], and could potentially do the same for vaccine-activated tumor-reactive T cells [168
].
IL-2
IL-2 activates STAT1 directly (although its major effects are mediated by STAT5 [171
]) and indirectly (by potentiating IFN- release from immune cells) [172
]. Signaling through IL-2Rs found on a number of hematopoietic [173
] and solid tumors [174
] has not been studied in complete detail and may sometimes promote tumor growth in vitro [174
]. However, use of IL-2 as a cancer vaccine adjuvant would likely also depend on its ability to activate cytotoxic effector cells and allow T cells to overcome the effects of cancer-derived immunosuppressive factors. For example, IL-2 could make T cells resistant to the suppressive effects of TGF-ß from cancer cells and Tregs [175
] by inhibiting nuclear translocation of SMAD2 and SMAD3 through IL-2-induced ERK-mediated phosphorylation in the T cells [176
]. IL-2 may also reverse T cell anergy [177
].
IL-2 is licensed as single-agent therapy for metastatic RCC and malignant melanoma and is being studied increasingly as a cancer vaccine adjuvant [178 ]. High-dose IL-2 was found to increase the therapeutic effectiveness of vaccines composed of HLA-A*0201-binding peptide epitopes of the melanoma antigen gp100 [179 ]. However, the optimal dose, route of administration, timing relative to vaccination, and use in combination with other immunomodulatory agents remain to be determined.
IL-12
IL-12 activates STAT4 directly and STAT1 indirectly by its ability to promote type 1 immunity [180
]. IL-12 is being studied as a vaccine adjuvant alone and in combination with IL-2 [181
].
IFN-
The most obvious, clinically relevant agent to activate STAT1-mediated signaling pathways and inhibit constitutive STAT3 activation in tumor cells (and T cells also) is IFN- itself. Indeed, injections of IFN-2b have been shown to decrease STAT3 expression in premalignant melanoma lesions [182
]. As with IL-2, IFN- has relatively modest clinical activity as a single agent in several tumors (such as RCC, CML, melanoma, and hairy cell leukemia), and the optimal method of administration is still not entirely clear [183
]. Recently, co-workers and I [5
] found that high doses of IFN-2b, administered for 1 month after a vaccine directed against gp100 (expressed in a Canary pox vector), produced striking immunological and therapeutic effects in patients with metastatic melanoma.
|
AMPLIFICATION OF VACCINE-INDUCED ANTI-TUMOR RESPONSES BY HDI |
|---|
75% of 24 HLA-A*0201 melanoma patients (Fig. 1A)
. As described in the Introduction, clinical and immunological responses were not maintained with the vaccines alone, and as some patients remained at high risk for disease progression (because of skin, lung, or lymph node metastases or deep primary lesions [184
]), intravenous IFN-2b (20x106 U/m2, 20 times over 4 weeks) was administered to seven of them at different times after the vaccine protocol [5
]. This regimen of HDI recalled gp100-reactive T cells in the four patients who had previously (but transiently) developed immunological responses to the vaccines (Fig. 1A) but not in the remaining patients in whom a previous response had not been detected. HDI was associated with disappearance of metastatic lesions (Fig. 1B) and acquisition of lytic function by gp100-reactive T cells (Fig. 1C) in two patients. These results suggested that HDI had "amplified" the effects of the vaccine by recalling previously activated, tumor-reactive T cells that were now able to mediate potent killing of melanoma cells.
The precise explanation for the increased number and potency of tumor-reactive T cells after HDI is not clear. NK and T cell function are increased in the blood of melanoma patients treated with HDI [185
]. However, the effects on T cells noted after vaccination (Fig. 1)
could not be reproduced in vitro by simply adding IFN- to cultures of PBMCs and gp100 antigens [5
], suggesting that direct effects of HDI on melanoma cells were also involved. In any case, this study provides a "proof of principle" in human subjects that agents, which counteract the gene programs of cancer cells that make them intrinsically resistant to tumor-reactive T cells, may improve significantly the results of cancer vaccines. The possible association of vaccination and HDI with STAT1 activation and STAT3 inhibition in tumor cells and T cells is being evaluated in further studies.
|
SUMMARY AND FUTURE DIRECTIONS |
|---|
Different cancers may depend disproportionately on the constitutive activation of one signaling pathway. For example, transforming RAS mutations are more prevalent in solid tumors [82
], and STAT3 activation predominates in some hematologic tumors [186
]. The dominant signaling pathway in a tumor can be indicated potentially by assessing the phosphorylation state of individual components with high-quality antibodies [88
] or the expression of characteristic genes with cDNA microarrays. This information could then be used to choose the most effective adjuvant for a patient. For example, IFN-2b may be more effective in cancers that are especially dependent on STAT3, and RAF inhibitors may be more effective where the ERK pathway dominates. Together with assays to assure that sufficient numbers of T cells have been activated appropriately after vaccination [187
], the ability to assess signaling in situ, using immunohistochemical or gene expression assays, could also be useful for evaluating treatment efficacy.
Combination therapy with a number of different immunomodulatory agents may be needed to achieve the desired effects on cancer cell gene programs. For example, IFN-2b may not be effective, where constitutive STAT3 activation has resulted in high SOCS1 levels in the cancer cells [188
]. However, the combination of IFN-2b (to activate STAT1) and AG490 (to inhibit STAT3) could make such cancer cells highly susceptible to tumor-reactive T cells. Although significant autoimmune pathology might accompany this approach, systemic activation of autoreactive T cells (if it occurs) should be self-limited and easily treated, as others and I [189
] have shown following syngeneic bone marrow transplantation.
|
ACKNOWLEDGEMENTS |
|---|
Received January 12, 2004; revised March 3, 2004; accepted March 9, 2004.
|
REFERENCES |
|---|
2b: implications for cancer vaccines Clin. Cancer Res. 9,4347-4355
TOP
ABSTRACT
INTRODUCTION