Soluble NKG2D ligands limit the efficacy of immune checkpoint blockade

The anticancer activity of CD8⁺ cytotoxic T lymphocytes (CTLs) is often maintained under control by inhibitory receptors including (but not limited to) cytotoxic T lymphocyte associated protein 4 (CTLA4) and programmed cell death 1 (PDCD1, best known as PD-1).¹ Demonstrating the prominent pathophysiological relevance of these immunological checkpoints, no less than 6 different monoclonal antibodies targeting CTLA4, PD-1 or its major ligand (CD274, best known as PD-L1) have been shown to mediate robust clinical activity and are now approved for use in patients affected by a variety of cancers.² Of note, the PD-1-targeting agent pembrolizumab (KEYTRUDA®, from Merck and Co.) has recently been granted accelerated regulatory approval by the US Food and Drug Administration (FDA) for the treatment of any solid tumor with elevated microsatellite instability³ or displaying a defect in mismatch repair⁴ (https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm560167.htm). This is the first time in history that a treatment is licensed for use in patients bearing any solid tumor with a specific genetic feature, irrespective of tissue of origin. Despite such an extraordinary clinical success, only a fraction of patients respond to and obtain therapeutic benefits from immune checkpoint blockers, owing to multiple mechanisms of resistance.⁵ For instance, the likelihood of metastatic melanoma patients to respond to pembrolizumab or the CTLA4-targeting antibody ipilimumab (YERVOY®, from Bristol-Myers Squibb) has been shown to decrease when tumors are characterized by a reduced neoantigenic burden⁶ or by defects in antigen presentation,⁷ as well as when the tumor immune infiltrate manifests functional defects in effector mechanisms.^6,8 Accumulating data suggest that the efficacy of immune checkpoint blockers is also compromised in the presence of soluble ligands for killer cell lectin like receptor K1 (KLRK1, best known as NKG2D).^9-11

NKG2D is a plasma membrane receptor expressed by CTLs and natural killer (NK) cells (as well as by other cells of the immune system) that normally triggers effector responses upon binding to a variety of ligands exposed on the surface of stressed, infected or malignant cells, including MHC class I polypeptide-related sequence B (MICB) and UL16 binding protein 1 (ULBP1).^12-14 However, NKG2D ligands (NKG2DLs) can also be shed from the surface of cancer cells upon metalloproteinase-mediated cleavage.^15,16 Such soluble NKG2DLs operate as decoy molecules and mediate powerful immunosuppressive functions that support the escape of malignant cells from immunosurveillance.^15,16 Accordingly, increased circulating levels of soluble NKG2DLs have been documented in patients with a variety of advanced cancers, correlating with a dismal prognosis.¹⁷

Zhang et al. set out to study the impact of soluble NKG2DLs on the efficacy of immune checkpoint blockers in transgenic adenocarcinoma of the mouse prostate (TRAMP) mice.¹¹ To this aim, they engineered TRAMP mice with a construct for the prostate-specific expression of human MICB, and investigated therapeutic responses to a murine CTLA4-targeting antibody. Increased levels of soluble MICB in the serum were associated with limited therapeutic efficacy, reduced survival and the development of autoimmune side effects including colitis. Consistently, the co-administration of a MICB-neutralizing antibody favored the expansion of tumor-specific CTLs and their recruitment to malignant lesions, resulting in empowered therapeutic responses to CTLA4 blockade and limited autoimmune complications from treatment.¹¹

Maccalli and colleagues performed a comprehensive phenotypic and functional immunomonitoring of circulating T cells and soluble immunomodulatory factors from melanoma patients receiving immune checkpoint blockers as standalone therapeutic agents or in the context of combinatorial treatment regimens.^9,10 In this setting, the absence of soluble MICB, ULBP1 or ULBP2 from the sera of patients before the initiation of treatment correlated with improved long-term survival. Moreover, a multivariate analysis of prognostic biomarkers attributed to the circulating levels of soluble ULBP1 and ULBP2 independent prognostic value for clinical outcome in melanoma patients receiving immune checkpoint blockers.^9,10

Collectively, these findings reveal the ability of soluble NKG2DLs to interfere with, and hence limit the therapeutic efficacy of, immune checkpoint blockers. Interestingly, the soluble variant of murine ULBP1 (also known as MULT1) mediates immunostimulatory, rather than immunosuppressive, functions.¹⁸ It will be interesting to study the role of this molecule (if any) in the anticancer immune response unleashed by immune checkpoint blockers. Additional investigation is also required to dissect the molecular and cellular mechanisms linking NKG2D signaling to immunological checkpoints in the context of cancer immunosurveillance. Nonetheless, NKG2DLs and the metalloproteases that are responsible for their shedding stand out as potential targets for the development of new strategies for cancer immunosensitization.