Blocking PD-1 in cancer immunotherapy

Abundant experimental evidence supports the role of the immune system in controlling tumor development and progression. Immunologic interventions including allogeneic hematopoietic stem cell transplantation, donor lymphocyte infusion, adoptive transfer of antigen-specific cytotoxic T lymphocytes (CTLs) and tumor/peptide vaccination have all shown clinical activity that supports the ability of the immune system to impair tumor growth and prevent disease recurrence. Conversely, tumor escape from immune recognition by down-regulation of MHC molecules and direct inhibition of the function of tumor-specific CTLs has also been extensively demonstrated. Indeed, the tumor environment exploits multiple factors that cooperate to inactivate immune responses, such as production of inhibitory cytokines (TGF-β), aberrant expression of proapoptotic T-cell ligands (Fas-ligand), production of tryptophan catabolic enzyme, indoleamine 2,3-dioxygenase (IDO), and accumulation of regulatory T cells.¹ 

Programmed cell death 1 (PD-1) receptor is rapidly up-regulated in T lymphocytes responding to viral infections, but under most circumstances it is quickly down-regulated upon removal of the antigen. By contrast, during chronic infections such as HIV²  and HCV,³  virus-specific CD8⁺ T lymphocytes show sustained expression of PD-1 which triggers T-cell dysfunction or exhaustion on interaction with antigen-presenting cells expressing PD ligand 1 (PD-L1).⁴  Previous reports demonstrated that the PD-1–PD-L1 pathway may also serve as a novel immune escape mechanism in mouse models of solid tumors.⁵  In this issue of Blood, 2 independent groups confirm the contribution of the PD-1 signaling pathway to immune tumor escape in 2 hematologic malignancies, chronic myelogenous leukemia (CML) and acute myelogenous leukemia (AML). Mumprecht and colleagues show that tumor-specific CTLs express high levels of PD-1 and have impaired function in a mouse model of CML.⁶  Importantly, these experiments confirm not only that PD-1 expression is a marker of T-cell exhaustion, but also that PD-L1 expressed by tumor cells contributes to T-lymphocyte dysfunction. Zhang and colleagues obtained similar results in a mouse model of AML, in which leukemic cells once again expressed PD-L1 in vivo, suggesting a direct pathogenetic role of the PD-1–PD-L1 pathway.⁷  In the third article, Ahamadzadeh and colleagues report an extensive analysis of melanoma samples and show that melanoma infiltrating T lymphocytes (TILs) consistently express PD-1 and have impaired functionality.⁸ 

The recognition that tumors generate a “hostile” microenvironment for antitumor effector cells has prompted investigators to develop strategies to overcome these barriers. Genetic modifications of tumor-specific CTLs to counter specific tumor-evasion strategies have been explored to enhance the efficacy of adoptive T-cell therapies. For example, tumor-specific CTLs have been engineered to become resistant to the inhibitory effects of TGFβ⁹  or the proapototic effects of Fas-ligand.¹⁰  A phase 1 clinical trial is ongoing in patients with Epstein-Barr virus (EBV)–related malignancies. Patients receive adoptively transferred EBV-specific CTLs genetically modified to express a dominant negative TGFβ receptor that confers resistance to the inhibitory effects of TGFβ. Clinical trials using Denileukin Diftitox to deplete in vivo regulatory T cells have also been reported,¹¹  although clinical benefit was not consistently observed.¹² 

The PD-1–PD-L1 pathway represents another challenging hurdle to be crossed if the clinical impact of immunotherapy of human malignancies is to be maximized. It is evident from complete responses observed in melanoma patients by Rosenberg's group that ex vivo expanded TILs can become functional effector T cells in vivo.¹³  Since disruption of PD-1–PD-L1 interactions by infusion of blocking antibodies enhances immune mediated antitumor responses in all mouse models so far reported (including those in this issue of Blood), it will be of great interest to learn whether PD-1–PD-L1 pathway inhibition in melanoma patients can further improve the success rate of TIL infusions. The development of clinical-grade reagents that interfere with this pathway should therefore be a priority.

The list of factors and mechanisms that tumors use to escape a destructive immune response grows ever longer, and it is evident that tumor cells can use more than one of these mechanisms to thwart immune attack. The question that remains is how many of these mechanisms need to be efficiently counter-attacked to allow the immune system to eradicate the tumor.