It has been known for many years that immunological mechanisms play a key role in cancers,1 including those which are most common in humans.2–5 However, for several decades, the benefit of active immunotherapy was modest. This situation is now rapidly changing. In particular, systemic administration of immunomodulatory monoclonal antibodies (mAbs) has demonstrated clinical efficacy, including complete remission in several patients treated at tolerable toxicity for melanoma or other particular types of tumours.6,7 These findings are remarkable but much remains to be learned.
The tumour microenvironment is highly immunosuppressive and local type 2 T helper (Th2)-type inflammation appears to play an important role in this immunosuppression.8–10 We hypothesized that shifting a Th2-type inflammation to a type 1 T helper (Th1) response at the tumour site will overcome immunosuppression and promote tumour destruction, and that combinations of immunomodulatory mAbs can be applied to achieve this. Initially, we used a combination of mAbs to CD137 (part of the tumour necrosis factor receptor family), programmed cell death protein 1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA4),11 selecting this combination because each of these mAbs displays therapeutic activity in some mouse tumour models, and also in some human patients. AntiCD137 mAb activates Th1-type CD8 and CD4 lymphocytes and natural killer cells and inhibits T cell-dependent antibody production. AntiPD-1 mAb counteracts immunosuppression induced by the PD-1/PDL-1 pathway and synergizes with antiCD137 mAb. AntiCTLA4 mAb counteracts immunosuppression by regulatory T cells (Tregs) and synergizes with antiPD-1 and antiCD137. Subsequently, we also added a mAb to CD19 (a B-lymphocyte antigen) in view of evidence that decreased numbers of CD19 cells in tumours and tumour-draining lymph nodes correlated with response to therapy,12 and a study was also carried out in collaboration with a group of investigators from Shanghai where, as part of the study, mice with established ovarian carcinoma were injected with cisplatin, one of the drugs most often used for chemotherapy of ovarian cancer, in addition to intraperitoneal injection of mAbs to CD137 and PD-1.
Four tumour models were investigated, the B16 melanoma, the SW1 clone of the K1735 melanoma, the TC1 lung carcinoma and the ID8 clone of the mouse ovarian surface epithelial cell ovarian carcinoma. The first three tumours were growing subcutaneously and had a mean diameter of 7–9 mm when therapy started. The ID8 tumour was transplanted intraperitoneally with treatment commencing 10 or 15 days after transplantation of 3 × 106 tumour cells. The mAbs were given either intratumorally or intraperitoneally to mice with subcutaneous tumours and intraperitoneally to mice with the ID8 tumour, with 500 or 250 μg of each mAb delivered weekly for 3 weeks and then biweekly for another 3 weeks. The mice were then followed for tumour growth and overall survival.
We have drawn the following main conclusions from our studies so far:
(1) Survival of mice with ID8 tumours tripled and > 50% of mice with B16, SW1 or TC1 tumours underwent long-lasting complete remission, which lasted over an observation period of > 150 days, and therefore were probably cured.11,12 It is noteworthy that the four-mAb combination (antiCD137/PD-1/CTLA4/CD19) also induced long-lasting remission when subcutaneous tumours had a mean diameter of 7–9 mm when treatment started,12 which is in contrast to findings obtained with other types of immunotherapy (e.g. therapeutic vaccination) where tumours of that size rarely responded. The three-mAb combination (antiCD137/PD-1/CTLA4) also exhibited powerful antitumour activity, curing some mice of their tumours, but it is significantly less efficacious than the four-mAb combination, particularly against large tumours. In the models we studied, single mAb treatment was ineffective, while a combination of mAbs to CD137 and PD-1 significantly prolonged survival when tested in the ID8 model.11,14 Our data illustrate the power of properly activated immunological mechanisms to destroy tumours and produce cures. It is noteworthy that the approaches used did not rapidly select for therapy-resistant cells and probably also killed tumour cells which have lost their ability to express or present tumour antigens, most likely via some bystander mechanism, as seen in some other systems.13 The observed toxicities were modest and primarily detected as temporary hair loss, depigmentation in a few cases for the B16 melanoma model and < 5% of death.
(2) Intratumoral injection of the mAb combinations is more efficacious than systemic (intraperitoneal) injection in mice with subcutaneous tumours and induces a stronger systemic antitumour response than systemic administration against a second, non-injected, tumour in mice with two tumours.12 The advantage of intratumoral over systemic injection is best seen in mice with large tumour loads or a second, not treated, tumour.
(3) The combination of cisplatin with mAbs to CD137 and PD-1 induces long-lasting tumour-free survival when given to mice with ID8 ovarian carcinoma established intraperitoneally.14 These findings are striking and suggest that an analogous approach should be seriously considered for patients with advanced ovarian carcinoma. There is precedent for giving therapeutic agents intraperitoneally to such patients, which probably will be more efficacious for the mAb combination than giving it systemically (although this needs to be investigated).
(4) Therapeutically efficacious approaches dramatically decrease tumour-infiltrating CD19 and CD20 cells, as well as such cells in tumour-draining and distal lymph nodes and spleen, including depletion of CD19+ cells which also express CD11c (an integrin alpha X protein), contain indoleamine 2,3-dioxygenase12 and are probably tolerogenic dendritic cells, as detected in other systems where they are localized primarily in the spleen.15 Furthermore, they decrease tumour-associated myeloid-derived suppressor cells and Treg cells and increase the numbers of tumour-associated CD4 and CD8 cells, as well as mature dendritic cells and long-term memory CD4 and CD8 T cells. The ID8 cells express a mouse homologue of the human ovarian cancer antigen mesothelin, and treatment with immunomodulatory mAb combinations induces an immune response to mouse mesothelin, as detected with both enzyme-linked immunospot and cytotoxic T-lymphocyte assays.
(5) We conclude that the mAb combinations we have studied, and particularly the four-mAb combination (antiCD137/PD-1/CTLA4/CD19), should be considered strong candidates for ‘translation’ to human cancer, particularly for tumours confined to a site that can be treated locally, such as carcinoma of the ovary, prostate, bladder and perhaps brain. However, toxicity needs be studied further since we expect that a combination of mAbs, which can dramatically change the immune response to tumours, may also induce autoimmunity and clinical application of the findings need to be pursued with caution. In particular, the observed liver toxicity with antiCD137 mAbs in humans is a concern that needs to be addressed, perhaps by using smaller doses of antiCD137 and delivering them locally.16 It is also possible that equally or more efficacious mAb combinations can be identified that have fewer side-effects or combinations between mAbs and chemotherapeutic drugs.