Exhaustion is getting complicated Free

Immune cell exhaustion is a hot-button issue in cancer therapy. In this issue of Blood, Myers et al¹ demonstrated that depending on the chronic activation signal given to natural killer (NK) cells, there are differential effects on an exhaustion phenotype. They demonstrated this in both mouse and human models through chronic activation of NK cells by triggering NK-cell activation receptors (NKARs) vs activation through cytokine alone. Continued exposure to interleukin-15 (IL-15) led to an impairment of NK-cell effector functions, but chronic NKAR triggering also suppressed NK chemotactic abilities, which affected tissue homing. Furthermore, mechanistically, chronic NKAR stimulation correlated with downregulation of the transcription factor, Krüppel-like factor 2 (KLF2). Using knockout mice, they demonstrated that a reduction in KLF2 mediates many of the exhaustion-like effects following prolonged NKAR signaling, also mirroring what occurs in the tumor microenvironment. Importantly, using a novel genetic knockin platform, this NK-cell impairment can be overridden.

Why is this important? First, it is well-recognized that immune cell exhaustion can be a major limiting factor in cancer immunotherapy, particularly demonstrated with T cells in which the use of immune checkpoint inhibition of targets like PD-1 has demonstrated marked clinical success. These advances have led to intense interest in discovering more checkpoint pathways to target. Recent efforts in NK-cell–based therapies, including chimeric antigen receptor NK cells and the observation that, like with T cells, similar exhaustion-like or anergy patterns can be exhibited,^2,3 have led to efforts to increase NK-cell function through immune checkpoint inhibition. However, applying exhaustion generically to chronically stimulated immune cells, whether they are T or NK cells, may be oversimplifying the processes. Classically, exhaustion refers to impaired function and proliferation of the effector cells as a consequence of chronic stimulation.⁴ The chemotaxis and homing effects have been far less characterized but may be of particular importance when applying adoptive cellular immunotherapy approaches, particularly with NK cells and hematologic malignancies. Second, the concept that chronic stimulation leads to an exhaustive state is often loosely and broadly applied to any stimulation. Both NKAR-triggering and cytokine-stimulation alone have been used to generate memory-like NK cells, suggesting that these cells may develop different properties depending on the activating stimuli that are received. The current study would suggest that important differences do exist, because the net result on the chronically stimulated NK cells differs, at least in terms of later impaired function.

When compared with T cells, NK cells readily respond to cytokine stimulation alone via IL-15 or IL-2, and in the current study, it enabled differentiation between chronic cytokine signaling and receptor triggering. Although impaired NK function has been observed under both conditions, only chronic NKAR triggering has led to chemotaxis and homing deficits, accompanied by a significant reduction in KLF2 mirroring, which has been observed within certain tumors and may be more of an issue with hematologic malignancies than with solid tissue cancers.

Of note, KLF2 has been determined previously to exert significant effects on T cells. The absence of KLF2 has been shown to produce unrestrained cytokine production and chemokine receptor upregulation in naïve T cells.⁵ Recently, however, using knockout mouse viral models, KLF2 also has been shown to maintain stem-like memory T cells, enable adequate recall responses, and, most revealingly, prevent exhaustion.⁶ In the present study by Myers et al, it seems that, yet again, similarities and differences exist between NK and T cells in terms of exhaustion. Using mouse models, Myers et al demonstrated that KLF2^–/– NK cells have defects in their ability to emigrate out of the bone marrow, the location where NK cells develop, in a situation analogous to the issue with naïve T cells. Similarly, at the tumor site where exhaustion occurs, KLF2 seems to be able to prevent these effects, as also observed with T cells. Given the constant turnover and generation of NK cells throughout life, this raises the question of whether KLF2 also plays a role in normal NK-cell development/differentiation or only during disease states. With NK cells, this study also demonstrated the need to assess potential effects in bone marrow, peripheral blood, and in target tissues. The role of KLF2 in cellular homing also possibly depends on the differentiation state of the NK cell because differential subset effects were not assessed.

With these findings, we must ask what the next steps are in NK-cell therapy. The present study is important because, in contrast with many earlier studies that assessed the loss of KLF2, Myers et al engineered a genetic knockin approach. This leads to critical follow-up studies to determine if engineering NK-cell products that overexpress KLF2 will lead to both increased tumor localization and effector function following adoptive transfer. Exploiting this pathway will be key but may have issues. A major limitation with NK cells resides in their heterogeneity and lack of clonal expansion properties, making it difficult to gauge the impact of genetic manipulation using whole populations. It will also be important to further delineate the pathway through which chronic NKAR triggering causes the reduction in KLF2 to determine other potential targets. NK-cell exhaustion effects have also been associated with other inhibitory molecules, such as TIGIT and Pik3ip1, that also need to be assessed in the context of KLF2.⁷ Finally, it will be important to determine if these heightened or prolonged activation thresholds affect NK-cell survival or differentiation. Nonetheless, the study by Myers et al now points to unraveling the different means through which NK cells inhibit themselves, how the method of inhibition differentially affects their functions, and how to possibly circumvent this. It also highlights a concept that has been acknowledged but not formally addressed in the cellular immunotherapy field, namely that the exhaustion process is more complex than originally thought when comparing NK and T cells, both in induction and in the different immunological processes affected.

Conflict-of-interest disclosure: The author declares no competing financial interests.