Although still early in clinical development, agonists of Toll-like receptor 9 (TLR9) have demonstrated potential for the treatment of cancer. TLR9 agonists directly induce activation and maturation of plasmacytoid dendritic cells and enhance differentiation of B cells into antibody-secreting plasma cells. Preclinical and early clinical data support the use of TLR9 agonists in patients with solid tumors and hematologic malignancies. In preclinical studies, TLR9 agonists have shown activity not only as monotherapy, but also in combination with multiple other therapies, including vaccines, antibodies, cellular therapies, other immunotherapies, antiangiogenic agents, radiotherapy, cryotherapy and some chemotherapies. Phase I and II clinical trials have indicated that these agents have antitumor activity as single agents and enhance the development of antitumor T-cell responses when used as therapeutic vaccine adjuvants. The activity and safety of these novel anticancer agents are being explored in a wide range of tumor types as part of a variety of therapeutic strategies with the goal of harnessing the immune response to fight cancer.
Toll-like receptor 9
The human immune system detects and responds to infectious challenges using several families of receptors that can recognize pathogen-expressed molecules, such as lipopolysaccharide, viral RNA or bacterial DNA (Akira et al., 2006). The most well understood of these receptors are the Toll-like receptors (TLRs), of which a family of 10 related molecules has been identified in humans (Akira et al., 2006). The immune role of TLR9 has been studied most extensively in plasmacytoid dendritic cells (pDCs) and B cells (Iwasaki and Medzhitov, 2004), which may be the only human immune cells to constitutively express TLR9. Cellular activation is reported to induce TLR9 expression in additional cell types, including human neutrophils (Hayashi et al., 2003), monocytes and monocyte-derived cells (Saikh et al., 2004; Siren et al., 2005) and CD4 T cells (Gelman et al., 2006), but the biologic role for this is less well understood. TLR9 expression has also been reported in some nonimmune cells, including pulmonary epithelial cells and lung cancers (Li et al., 2004; Platz et al., 2004; Droemann et al., 2005), keratinocytes (Lebre et al., 2007) and intestinal epithelium (Pedersen et al., 2005; Lee et al., 2006).
TLR9 recognizes and is activated by unmethylated cytosine-phosphate-guanine (CpG) dinucleotides, which are relatively common in bacterial and viral DNA but are suppressed and methylated in vertebrate DNA (Krieg, 2004). Binding of DNA containing unmethylated CpG motifs to TLR9 causes a conformational shift in the receptor, which is thought to result in recruitment of the adapter protein MyD88, activation of signaling pathways with the phosphorylation of mitogen-activated protein kinases and activation of nuclear factor-κB (Latz et al., 2007). At a cellular level, activation of TLR9 initiates a cascade of innate and adaptive immune responses (Figure 1). TLR9 agonists activate pDCs to secrete type I interferon (IFN) and to express increased levels of co-stimulatory molecules such as CD80 (B7.1) and CD86 (B7.2). This is believed to initiate a range of secondary effects, including secretion of cytokines/chemokines (for example, monocyte chemoattractant protein-1 (MCP-1), IFN-γ-inducible 10 kDa protein (IP-10)), activation of natural killer (NK) cells and expansion of T-cell populations, particularly type 1 helper T (TH1) cells and cytotoxic T lymphocytes (CTLs) (Krieg, 2004, 2006). As a result a potent, cell-mediated TH1 response is initiated. A humoral immune response is also initiated as TLR9 agonists enhance differentiation of B cells into antibody-secreting plasma cells, potentially promoting antibody-dependent cellular cytotoxicity (Appay et al., 2006). An understanding of the immune cascade initiated by TLR9 activation has prompted the clinical development of several TLR9 agonists in the fields of infectious disease, cancer and asthma/allergy. The rationale for investigating TLR9 agonists as anticancer agents is based on the hypothesis that the innate immune activation may have direct antitumor effects and that the enhanced tumor antigen presentation in a TH1-like cytokine and chemokine milieu will promote an antitumor immune response.
Several synthetic oligodeoxynucleotide (ODN) agonists for TLR9 are currently in development for the treatment of cancer. Because the phosphodiester bond of native DNA is rapidly degraded by endonucleases, these investigational CpG ODNs use a nuclease-resistant phosphorothioate backbone that improves the half-life in the body from just a few minutes (for unmodified native DNA) to about 48 h. Coley Pharmaceutical Group (Wellesley, MA, USA) has developed a CpG ODN known as CPG 7909 that is being investigated as a vaccine adjuvant in several tumor types (Table 1). CPG 7909 was licensed by Pfizer Pharmaceuticals Inc. (New York, NY, USA) for clinical investigation as a single agent or in combination with other therapeutic approaches, and is now known as PF-3512676 when it is not being used as a vaccine adjuvant. Other TLR9 agonists in clinical development for cancer include ISS 1018 (Dynavax Technologies, Berkeley, CA, USA), IMO-2055 (Idera Pharmaceuticals Inc., Cambridge, MA, USA) and CpG-28 (University of Paris, France).
Cancer therapy with TLR9 agonists
The initial impetus to develop TLR9 agonists as anticancer drugs came from several preclinical studies demonstrating antitumor activity in a wide variety of tumor models. For example, in a murine cervical carcinoma model, mice with established subcutaneous (s.c.) tumors treated with CpG ODNs injected at a distant site exhibited significant regression (P<0.05), and treated mice had a significant improvement in survival (P<0.001) compared with control mice (Baines and Celis, 2003). In addition, >50% of mice (24 of 42) with established tumors had a complete regression following treatment for 9 days (Baines and Celis, 2003). TLR9 agonists also have been successfully combined with traditional anticancer therapies (for example, radiation therapy, chemotherapy) or other immunotherapies (for example, vaccines). For example, in mice with established orthotopic rhabdomyosarcoma tumors, treatment with a TLR9 agonist plus topotecan resulted in an improvement in survival rate (41 versus 22% for topotecan alone, P=0.09) (Weigel et al., 2003). The combination of a CpG ODN with radiation therapy also enhanced tumor response in a murine fibrosarcoma model (Milas et al., 2004). However, mice and humans have a different TLR9 expression pattern, and exposure to CpG motifs stimulates a more narrow profile of cytokines/chemokines in humans (Krieg, 2007). Thus, preclinical results cannot be considered predictive of clinical findings. However, based on the strength of the extensive preclinical data, several TLR9 agonists are now being developed as anticancer agents.
Monotherapy with TLR9 agonists
In addition to enhancing maturation of B cells into antibody-secreting plasma cells, TLR9 agonists have a variety of other effects on B cells that may be relevant in the treatment of hematologic malignancies. Activation of TLR9 on primary malignant B cells upregulates expression of major histocompatibility complex molecules and other surface receptors, thereby increasing their capacity to stimulate T cells (Decker et al., 2000; Jahrsdorfer et al., 2001; Wooldridge and Weiner, 2003). This may result in an enhanced T-cell-mediated response to tumor antigens on the malignant B cells. Furthermore, in patients with advanced cutaneous T-cell lymphoma (CTCL), NK cells and CD8+ T cells were activated following culturing of peripheral blood mononuclear cells with CpG ODNs, and there was a marked increase in IFN-α production (Wysocka et al., 2004). This could enhance the antitumor immune response by activating NK cells, or may induce a direct antiproliferative effect (Sabel and Sondak, 2003). Chronic lymphocytic leukemia (CLL) cells, which express TLR9, are induced by CpG ODN to undergo apoptosis, in contrast with normal primary B cells in which TLR9 activation protects against apoptosis (Jahrsdorfer et al., 2005).
Several clinical studies of single-agent TLR9 agonists have been completed in patients with hematologic malignancies. In a phase I study of PF-3512676 in patients (N=28) with CTCL, seven (25%) patients achieved an objective response as determined by a Physician Global Assessment (two complete responses (CRs), five partial responses (PRs)) (Kim et al., 2004). Common adverse events (AEs) included mild to moderate injection-site reactions (erythema, induration, edema, inflammation and pain) and flu-like symptoms (fatigue, rigors, fever, arthralgia) (Kim et al., 2004). PF-3512676 has also been studied in a phase I trial in patients with refractory non-Hodgkin's lymphoma (NHL). In patients (N=23) receiving PF-3512676 (0.01–0.64 mg kg−1) intravenously (i.v.) up to three times a week, NK cell numbers and activation were enhanced in most subjects (Link et al., 2006). The majority of AEs were mild to moderate and transient. Serious hematologic AEs included anemia, thrombocytopenia and neutropenia, but were mainly judged to be due to disease progression (Link et al., 2006). In both trials, PF-3512676 was determined to be well tolerated at levels that can stimulate immune activation. These studies indicated that TLR9 agonists may be useful in the treatment of lymphoma. A phase I study is currently being carried out with PF-3512676 as second-line treatment for patients with CLL.
Because melanoma is a highly immunogenic tumor that has been shown to respond to immunotherapy, it is a logical malignancy in which to explore the activity of TLR9 agonists. Two general approaches have been investigated using monotherapy with PF-3512676: local therapy with intra- or perilesional injection and systemic therapy with s.c. injection. PF-3512676 demonstrated clinical activity in a phase I trial of intra- or perilesional injection in patients with metastatic melanoma (n=5, with one local regression) or basal cell carcinoma (n=5; with one CR and three PRs) (Trefzer et al., 2002; Hofmann et al., submitted). Treatment was associated with increased levels of serum interleukin (IL)-6, IL-12p40 and IP-10 in some patients, and cellular infiltrates of CD8+ lymphocytes were found after treatment in most lesions. In a different trial involving patients with clinical stage I/II melanoma, surgical resection of the primary tumor was followed by randomization to receive either saline or PF-3512676 (8 mg) intradermally at the excision site, followed 1 week later by a sentinel lymph node procedure (Molenkamp et al., 2007). In this setting, PF-3512676 was shown to induce the release of inflammatory cytokines and decrease the number of regulatory T cells observed in sentinel lymph nodes of patients with stage I to III melanoma (Molenkamp et al., 2007). Furthermore, both myeloid and pDCs were activated (Molenkamp et al., 2007), indicating that PF-3512676 clearly has immunomodulatory activity. In a different phase II study in patients (N=20) with metastatic melanoma treated with s.c. PF-3512676, two (10%) patients had a PR and three (15%) patients had stable disease (SD) (Pashenkov et al., 2006). These patients with possible clinical benefit from PF-3512676 therapy tended to have increased levels of NK cell activity compared to the patients with progressive disease. The majority of AEs were mild, consisting of short-term local injection-site reactions and transient flu-like symptoms (2 days); grade 3/4 laboratory AEs typically resolved without intervention. These studies demonstrated that PF-3512676 is generally well tolerated as a single agent in patients with melanoma and is associated with antitumor activity.
Other solid tumors
Monotherapy with TLR9 agonists has also been evaluated in phase I trials in patients with advanced renal cell carcinoma (RCC) or recurrent glioblastoma. A minor response was observed in 2 of 24 patients with recurrent glioblastoma receiving CpG-28 in a phase I trial. Treatment-related AEs consisted mainly of fever (21%), a worsening of neurologic conditions (17%) and reversible grade 3 lymphopenia (29%) (Carpentier et al., 2006). In a phase I dose-escalation trial in patients (N=31) with advanced RCC receiving PF-3512676, one patient (3%) had a PR and nine patients (29%) had SD (Thompson et al., 2004). No treatment-related serious AEs were observed, and injection-site reactions and flu-like symptoms were dose related and transient (Thompson et al., 2004). A phase I trial is currently investigating the activity of IMO-2055 as first- and second-line treatment of patients with RCC.
Combination therapy with TLR9 agonists
Preclinical studies have indicated that the combination of a TLR9 agonist and rituximab (Rituxan), a monoclonal antibody (mAb) against CD20, may be more effective than rituximab alone. In a murine lymphoma model, the addition of a CpG ODN to antitumor mAb therapy (with an anti-idiotype Ab) reduced the percentage of mice developing a tumor by 70% (Wooldridge et al., 1997). Two phase I clinical trials have investigated the combination of a CpG ODN with rituximab in patients with relapsed/refractory NHL. In a dose-escalation study, 50 patients received PF-3512676 weekly for 4 weeks i.v. or s.c. in combination with rituximab (375 mg m−2 per week × 4 weeks) i.v., while an additional s.c. cohort received weekly monotherapy with PF-3512676 for an additional 20 weeks following the initial 4 weeks of treatment with rituximab (Leonard et al., 2007, in press). Twelve patients (24%) had an objective response (five CRs, seven PRs), including 50% of the 12 patients receiving extended dosing with PF-3512676 (two CRs, four PRs) (Leonard et al., 2007, in press). The majority (76%) of patients experienced mild to moderate treatment-related AEs (primarily injection-site reactions and systemic flu-like symptoms). Few grade 3/4 AEs occurred in more than one patient or at more than one dose level, but four patients developed grade 3/4 neutropenia (Leonard et al., 2007, in press). In a phase I study, 20 patients with relapsed NHL received ISS 1018 weekly for 4 weeks s.c. following the second of four infusions with rituximab (375 mg m−2 per week × 4 weeks) (Friedberg et al., 2005). The overall response rate was 32% (one unconfirmed CR, five PRs). Common AEs included mild to moderate injection-site reactions and systemic flu-like symptoms (Friedberg et al., 2005). Combination therapy with TLR9 agonists is being further evaluated for antitumor activity in a phase I/II trial in combination with radiation therapy and in a phase II trial in combination with rituximab.
Non-small cell lung cancer
Several preclinical models have also suggested that a TLR9 agonist can synergize with cytotoxic chemotherapy (reviewed in Krieg, 2007). In a Lewis lung cancer model, the combination of PF-3512676 and paclitaxel significantly enhanced median survival compared with control mice (P<0.0001), and no additive toxicity was observed (Weeratna et al., 2004). Based on both preclinical and clinical proof-of-concept studies, a randomized controlled phase II trial was conducted in chemotherapy-naive patients with stage IIIb/IV non-small cell lung cancer (NSCLC) (Manegold et al., 2007, in press). Patients received 4–6 cycles of taxane/platinum chemotherapy with 0.20 mg kg−1 PF-3512676 s.c. on weeks 2 and 3 of each cycle (n=75) or chemotherapy alone (n=37). The objective response rate was 38% in the PF-3512676 plus chemotherapy arm (n=74) versus 19% in the chemotherapy-alone arm (n=37) by investigator evaluation, and the median overall survival was 12.3 months in the PF-3512676 plus chemotherapy arm versus 6.8 months in the chemotherapy-alone arm (hazard ratio=0.747, P=0.188) (Manegold et al., 2007, in press). The most common PF-3512676-related AEs were mild to moderate injection-site reactions and flu-like symptoms, although PF-3512676 was associated with an increase in the rate of neutropenia, anemia and thrombocytopenia.
Following completion of the phase II trials, two phase III studies were conducted exploring the combination of PF-3512676 (0.2 mg kg−1 s.c. days 8 and 15) with paclitaxel/carboplatin (200 mg m−2 per area under the curve six i.v. day 1) or gemcitabine/cisplatin (1250 mg m−2 i.v. day 1 and 8/75 mg m−2 i.v. day 1) versus chemotherapy alone as first-line treatment of patients with advanced NSCLC (Readett et al., 2007). No improvement in overall survival or progression-free survival was observed when PF-3512676 was added to standard platinum-based doublet chemotherapy in either trial. However, there was a higher frequency of grade 3/4 neutropenia and thrombocytopenia and a higher frequency of ‘sepsis-like events’ and ‘septic deaths’ reported as serious AEs in the PF-3512676 plus chemotherapy arms (Readett et al., 2007). Based on the recommendation of an independent data monitoring safety committee, these phase III studies were terminated, along with two phase II trials in which PF-3512676 was also combined with cytotoxic chemotherapy in patients with advanced NSCLC.
TLR9 agonists as vaccine adjuvants
A distinguishing characteristic of CpG ODNs is their ability to induce strong CD4+ and CD8+ T-cell responses and rapid production of antigen-specific antibodies when used as a vaccine adjuvant with many types of antigen (Krieg, 2007). These CpG ODN adjuvants have been heavily studied for the treatment of infectious diseases, and several clinical trials are also investigating their activity as tumor vaccines. The addition of CPG 7909 to a peptide MART-1 vaccine in patients with human leukocyte antigen-A2+ melanoma caused an approximate 10-fold increase in the number of antigen-specific CD8+ T cells (Speiser et al., 2005; Appay et al., 2006). Anti-MAGE-3 antibody titers were enhanced with the addition of CPG 7909 to a MAGE-3 recombinant protein vaccine in a phase I/II trial, and one patient had a durable objective response (PR for 9+ months) (van Ojik et al., 2002). CPG 7909 is also being studied as a vaccine adjuvant in patients with metastatic RCC. Preliminary results suggest that vaccination with autologous tumor cells, CPG 7909 and granulocyte macrophage colony-stimulating factor followed by maintenance therapy with IFN-α and PF-3512676 has antitumor activity in this population (3 of 12 patients had a PR) (Van Den Eertwegh et al., 2006). Phase I and II trials have been further investigating CPG 7909 as a vaccine adjuvant in patients with melanoma and breast cancer. The most advanced cancer vaccine program is a phase III, randomized, controlled, clinical trial of the tumor antigen MAGE-A3 combined with CPG 7909 (and other adjuvants) for the treatment of patients with early-stage (stage IB, II or IIIA), completely resected NSCLC whose tumors express the antigen (Table 1).
Safety of TLR9 agonists
Overall, TLR9 agonists are generally well tolerated in patients with cancer. The most common AEs observed with administration of TLR9 agonists are local injection-site reactions (for example, erythema, edema, inflammation and pain) or systemic flu-like symptoms (for example, headache, rigors, pyrexia, nausea and vomiting). These symptoms typically develop with 24 h of dosing and are transient, generally lasting for less than 2 days (Krieg, 2006). Grade 3/4 hematologic AEs (that is, anemia, neutropenia and thrombocytopenia) have been observed with TLR9 agonists (Link et al., 2006; Leonard et al., 2007, in press; Manegold et al., 2007, in press), and may require monitoring in future clinical trials. However, unlike other therapeutic immunotherapies for cancer such as IFN-α, TLR9 agonist therapy has not been associated with clinically significant autoimmune diseases.
Although it is still early in their clinical development, TLR9 agonists are generally well tolerated and have demonstrated potential as a novel immunotherapeutic approach to the treatment of cancer. Objective tumor responses to TLR9 agonist monotherapy have been observed in multiple tumor types, including both solid tumors and hematologic malignancies. Although these results are encouraging evidence of the potential clinical benefits of TLR9 activation, the greatest improvement in patient outcomes is likely to result from the use of this approach in combination with other therapies that work in a synergistic manner. Preclinical work and early clinical trials suggested that combination with cytotoxic chemotherapy may be one such approach, but two phase III trials of PF-3512676 administered concurrently with chemotherapy for patients with NSCLC have recently reported a lack of incremental efficacy with the addition of PF-3512676. The reason for the failure of these trials is unclear, but possible factors include the advanced stage of the disease, the use of concomitant steroid therapy as premedication for the chemotherapy, and possible immune-suppressive effects of the repeated chemotherapy cycles. Therapies based on activating an immune response against cancer may be expected to have a higher chance for success in clinical settings of earlier disease, such as the ongoing NSCLC vaccine trial in the adjuvant setting.
Combination therapy with antiangiogenic agents should avoid the immune-suppressive effects of chemotherapy and also has shown benefit in preclinical models when combined with TLR9 agonists (Damiano et al., 2006, 2007). An ongoing clinical trial in NSCLC is addressing the safety and potential clinical benefit of PF-3512676 combined with the antiepidermal growth factor receptor agent erlotinib.
Tumors can develop a variety of mechanisms to evade or suppress the immune system, and combination immunotherapeutic approaches may allow for multiple opportunities to slow or halt tumor progression. T-cell expression of CTL-associated antigen 4 (CTLA4) is induced during immune responses as a brake on the immune system, reducing the magnitude and functional capacity of the effector T-cell response. Blockade of CTLA4 with mAbs sustains T-cell proliferation and activation. By stimulating the DC side of the immune response and enhancing antigen presentation to T cells, TLR9 agonists may synergize with the effect of anti-CTLA4 mAbs ability to remove the brake on the T cells. Indeed, in preclinical tumor vaccine models, the combination of a TLR9 agonist and CTLA4 blockade led to improved rejection of established tumors through complementary mechanisms: the effect of anti-CTLA4 mAbs is largely dependent on CD4+ T cells, whereas the effect of the combination appears to be mediated by CD8+ T cells (Davila et al., 2003; Daftarian et al., 2004). The combination of anti-CTLA4 mAb and PF-3512676 is currently being evaluated in a phase I, open-label, nonrandomized, dose-escalation study in patients with metastatic melanoma and may provide a foundation for future therapeutic approaches. At the present time, the use of TLR9 agonists for enhancing tumor vaccination is the most advanced application of this technology in oncology, and the early results from this approach appear promising. Clinical trials are ongoing in other settings and in combination with other therapeutic approaches, and the true therapeutic potential of TLR9 agonists for the treatment of cancer remains to be determined.
Financial support for medical editorial assistance was provided by Pfizer Pharmaceuticals. I thank Todd Parker, PhD, ProEd Communications Inc., for expert editorial assistance with manuscript preparation.
About this article
Conflict of Interest
Arthur M Krieg is a founder, employee and shareholder in Coley Pharmaceutical Group and has a financial interest in the development of TLR9 agonists for cancer therapy.
Cellular and Molecular Life Sciences (2018)