Weal and woe of interleukin-18 in the T cell therapy of cancer

Abstract

Chimeric antigen receptor (CAR) T cell therapy of solid cancer remains below expectations; adding cytokine help through IL-18 has shown remarkable efficacy in first clinical trials. As IL-18 is also a powerful driver of hyperinflammatory conditions, we discuss to what extent unleashing IL-18 is a double-edged sword in CAR T cell therapies.

Introduction

Inducing and sustaining T cell-mediated immune responses against tumor-associated targets is key for successful cancer therapy in the long term. Immunotherapies with immune checkpoint inhibitors created plateaus of relapse-free survival in patients with chemorefractory disease revealing the potential of T cells to control cancer. Since this strategy relies on the presence of endogenous cancer-specific T cells, efficacy is limited to a minority of patients with cancer. Adoptive transfer of T cells with ex vivo engineered chimeric antigen receptors (CARs) with specificity for cancer-associated antigens allows a broader application and has shown efficacious in eradicating lymphoid malignancies.¹ However, it is producing limited activity against solid tumors, making it necessary to enhance T cell cytotoxicity while making the tumor microenvironment more conducive to immune destruction.

Cancer immunologist view

Solid tumors actively and passively prevent their immune destruction by creating a hostile environment, including the release of immune repressive cytokines and metabolites, deprivation from nutrients, low pH and strong physical barriers. While T cells, macrophages and NK cells are infiltrating many tumors, these cells are converted into an inactive, dysfunctional state that sustains rather than prevents tumor progression. A novel strategy is needed to repolarize infiltrated immune cells and to change the environment that allows adoptively transferred, infiltrating T cells to execute their antitumor response. In this context, interleukin-18 (IL-18), formerly named IFN-γ-inducing factor, has the potential to strongly activate T cells and NKG2D⁺ NK cells, convert CD8⁺ T cells into FoxO1^low Tbet^high killer cells, increase the release of proinflammatory cytokines IFN-γ, IL-22, IL-27, and repolarize suppressive macrophages toward an anticancer response.^{2 3}

Oncologist view

IL-18 is a pleiotropic cytokine impacting the overall immune response situation as it globally activates cytolytic T cells and NK cells supporting proliferation, differentiation and/or effector molecule expression by CD4⁺ helper T cells (T_H cells), mucosal-associated invariant T (MAIT) cells and γδ T cells.⁴ As a member of the IL-1 cytokine family, IL-18 shares many biochemical and immunological features with other cytokines like IL-1β, including structural similarities, the proteolytic cleavage of the pro-form to generate the signaling-competent cytokine, and the mode of release. Further, the level of signaling IL18 is controlled in a tightly orchestrated fashion by the IL-18 binding protein (IL-18bp) that acts as decoy-receptor (figure 1).⁴ IL-18 also orchestrates the viral defense by activating cytotoxic T cells to ultimately mediate pathogen clearance.⁵ Taken together, the potency to activate cytotoxic immune cells in an inflammatory environment renders IL-18 attractive for boosting current anticancer immune therapeutics.^{2 3} Such can be achieved by promoting endogenous IL-18 expression, by liberating IL-18, produced by tissue-resident macrophages, from IL-18bp using an IL18bp blocker,⁶ or by engineering CAR T cells for inducible or constitutive IL-18 release.⁷

Figure 1. Interleukin 18—a double-edged sword in modern cell therapy. Inflammatory challenge and/or viral infection¹ can trigger interleukin (IL) 18 expression and release from, that is, myeloid cells² which can then induce T cell (CD8, NK, CD4) activation/expansion and IFNγ expression. In turn, IFNγ-signaling to lymphocytes or endothelial and epithelial cells³ can induce expression of IL-18 binding protein (IL-18bp, 4) as the endogenous inhibitor of IL-18, thus quenching an IL-18-driven immune response. An imbalance in IL-18bp and free/unbound IL-18 favors T cell activation and associates with hyperinflammation in context of hemophagocytic lymphohistiocytosis or macrophage activation syndrome (MAS). While recent therapeutic approaches in context of treatment-refractory and/or solid tumors strive for enhancing IL-18 bioactivity by blocking its depletion by IL-18bp, or aim at boosting cell therapy by targeted IL-18 overexpression, in (auto)inflammatory conditions and associated MAS, treatment rather aims at a shut-down of excessive IL-18 expression and its systemic depletion to control hyperinflammation. DR IL-18: decoy-resistant IL-18; JAKi: JAK-inhibitor; MAS, macrophage activation syndrome; mTORi: mechanistic Target of Rapamycin inhibitor; rhIL-18bp: recombinant human IL-18bp.

Cancer immunologist view

To avoid the IL-18 impact on the overall immune response, CAR T cells were genetically engineered to release IL-18 in a locally defined fashion or to express membrane anchored IL-18, thus enhancing the antitumor activity in a cell-restricted fashion while reducing the risk of systemic effects. CAR T cells secreting IL-18 in a constitutive or antigen-triggered fashion, gained superior activities against solid tumors compared with standard CAR T cells in preclinical models,8,10 again, without increase in IL-18 serum levels.

Rheumatologist view

While oncologists strive for a sustained and targeted redirection of T cell responses, in many (auto)inflammatory or autoimmune conditions treatment aims at a shut-down of B cell and/or T cell activation to dampen effector molecule release. In such lines, targeted B cell depletion has resulted in some beneficial outcome in several autoimmune settings, but in some conditions, disease remitting effects were only weak or transient.¹¹ More recently, the introduction of CAR T cell treatment for refractory B cell autoimmune conditions demonstrated a more complete and durable re-set of a misdirected immune response by the targeted elimination of B cells,^{12 13} likely due to a more effective B cell pan-lineage depletion, including bone marrow resident cells.^{14 15}

IL-18 is a key cytokine in orchestrating the immune defense against viral infections but can also severely aggravate inflammatory conditions. For instance, in settings with fulminant, persisting viral load or an infectious scenario on the background of some inflammatory and rheumatologic diseases, such as Still’s disease (systemic juvenile idiopathic arthritis (sJIA); adult-onset Still’s disease (AOSD)), Kawasaki disease, or systemic lupus erythematosus, IL-18 can drive a potentially fatal cytokine storm known as secondary (infection-associated) hemophagocytic lymphohistiocytosis (sHLH, ia-HLH) or (i.e. sJIA-associated) macrophage activation syndrome (MAS).^{16 17} In mice, a condition recapitulating sHLH/MAS can be induced on repetitive injection of CpG DNA (TLR-9 ligand) to mimic a viral infection, or by combined application of viral mimetics, like p:IC (TLR3-ligand), and inflammatory challenge such as LPS (TLR4-ligand).^{18 19} Importantly, in sHLH and particularly sJIA-associated MAS, a massive IL-18 overexpression has the capacity to outnumber IL-18bp,^{17 20} thereby overrunning the physiological balance. As a result, unopposed IL-18 primes and feeds a vicious hyperinflammatory loop consisting of T cell IFN-γ release, IFN-γ-driven macrophage/myeloid cell activation and IL-18-driven T cell cytotoxicity (figure 1).^{16 17}

CAR T cell therapy can per se, without additional IL-18 support, produce life-threatening inflammatory complications. These include high-grade cytokine release syndrome (CRS), a hyperinflammatory condition as a consequence of CAR T cell expansion with some immunophenotypic overlap with sHLH/MAS,²¹ and late-onset inflammatory toxicity closely resembling sHLH/MAS and named immune-effector cell associated HLH-like syndrome (IEC-HS). Serum proteomics profiling in patients receiving CD19 CAR T cell therapy identified IL-18 to best recapitulate the onset of immune effector-associated neurotoxicity syndrome (ICANS),²² or to link with prolonged inflammatory hyperferritinemic toxicity and severe and persistent pancytopenia as recently demonstrated for CD37-specific CAR T cells (ClinicalTrials.gov/NCT04136275).²³ Increases in IFN-γ and total IL-18 (free IL-18 plus IL-18:IL-18bp complexes) were also reported after treatment with CD22-specific CAR T cells (NCT02315612). Comparisons of patients developing IEC-HS versus CRS suggested a disruption of the IL-18:IL-18bp equilibrium as a mediator of secondary inflammation.²⁴

Thus, an enforced IL-18 expression by engineered CAR T cells in a setting which is vulnerable to hyperinflammation may further increase the risk for severe cytokine storm conditions and should be handled with care.

Oncologist view

The physiological inflammatory homeostasis based on the IL-18:IL-18bp equilibrium deserves specific attention in the context of hyperinflammation and CAR T cell therapy-associated complications. Historically, the development of IL-18-based treatments started with recombinant IL-18 based on its capacity to control tumor growth in preclinical murine models.25,27 However, clinical applications of recombinant IL-18 in cancer patients, while safe and tolerated, disappointingly revealed limited efficacy,²⁸ likely due to IL-18bp which effectively neutralizes both the administered as well as endogenous IL-18.³

In fact, the high levels of IL-18bp in the TME not only protect from inflammatory toxicities but also prevent an IL-18-mediated antitumor effect. Such observations stimulated the development of decoy-resistant (DR) IL-18 variants³ or high-affinity anti-human IL-18bp monoclonal antibodies to reduce endogenous IL-18bp levels and/or to disrupt the IL-18:IL-18bp heterodimer to liberate IL-18 (figure 1). Both approaches have been successfully tested in preclinical models and in combination with immune checkpoint inhibitors or IL-12^{6 29} and are investigated in clinical trials, for example, NCT04787042.

Despite a key role for free IL-18 to drive cytokine storm conditions, no evidence pointed for an elevated risk of hyperinflammatory events in preclinical scenarios.^{6 29} Of note, transgenic IL-18 overexpression or genetic deletion of IL-18bp did not result in any spontaneous (hyper)inflammatory responses in mice.^{16 17}

Rheumatologist view

The lack of spontaneous inflammation in IL-18bp knock-out mice remains indeed an intriguing observation. However, when simulating a viral insult in Il18bp^−/− animals by repetitive CpG DNA injection, the IL-18bp knock-out enhanced susceptibility for hyperinflammation and aggravated the clinical course.¹⁶ More important, excess transgenic IL-18 combined with perforin deficiency, either homozygous or heterozygous, resulted in spontaneous and fatal IFN-γ-driven hyperinflammation.^{30 31} This observation is of seminal importance, as it recapitulates a frequent scenario in sJIA-associated MAS where up to every third patient (36%) has at least one heterozygous variant affecting the perforin pathway.³² Such variants, canonically associated with primary or familial HLH, lower the threshold for developing hyperinflammation. In combination with IL-18 overexpression on viral infection, these data highlight a dangerous interaction of even mild cytolytic defects with IL-18 excess.^{30 31 33}

Oncologist view

When using treatment with CAR T cells engineered for IL-18 expression, the mentioned scenarios argue for a high level of awareness and a personalized risk management. Functional screening for primary/familial HLH variants is well established in clinical diagnostics. Should cases of high-grade late-onset inflammatory toxicity be clinically observed, a potential association with preexisting variants should be addressed. Screening patients for germline variants prior to enrolment may be considered.

Meanwhile, initial clinical testing of safety, dose-escalation with a CD19-specific CAR T cell product with constitutive IL-18 secretion in adult patients with refractory non-Hodgkin lymphomas has been well tolerated. Notably, low cell doses achieved responses in patients who had relapsed after previous standard CAR T cell therapy.⁷ No high-grade or late-onset inflammatory toxicities were observed.⁷ These early data encourage further clinical studies using IL-18 augmented CAR T cell products.

Cancer immunologist view

On top of these initial and encouraging data, several current developments are aiming at local delivery of IL-18.^{8 9} This has two effects: a more focused treatment of solid tumors where CAR T cell activation occurs and the avoidance of life-threatening systemic inflammatory toxicities. IL-18 accumulation in tumor tissues may more likely reach the activation threshold for infiltrating adaptive and innate immune cells to enhance their antitumor activity (figure 1).

Joint perspective

Leveraging IL-18 in antitumor therapies for activating both targeted T cells as well as innate cells appears promising; however, the first clinical explorations may require close monitoring and an appropriate risk management strategy. Molecules and pathways that act as triggering drivers in the pathogenesis of hyperinflammation, on either an inflammatory disease, infectious or immune cell therapy background, can evolve as selective and druggable targets for salvage therapies, if required. As an exemplary hallmark, the IL-6R antagonist tocilizumab effectively rescues patients with life-threatening CRS on CAR T cell treatment.³⁴

While previous IL-18 applications in cancer treatment were hampered by the endogenous IL-18bp and thus never reached the threshold to drive hyperinflammatory side effects, the situation may change with pharmacological disruption of IL-18bp binding or with CAR T cells engineered for IL-18 expression. In this respect, mechanistic understanding of sHLH/MAS pathology can inform about targeted anti-inflammatory interventions beyond standard therapy with corticosteroids and/or etoposide as in primary HLH. Examples include the recombinant IL-1 receptor antagonist anakinra³⁵ and the antagonistic IFN-γ antibody emapalumab,³⁶ both have successfully been adopted for CAR T cell-induced CRS and IEC-HLH.37,39 Vice versa, following the discovery that tocilizumab can effectively abrogate CRS post-CAR T cell therapy, investigations regarding the effect of IL-6R blockade are initiated in sHLH patients.⁴⁰

To control potential sHLH/MAS-like scenarios after IL-18 CAR-T cell therapies, emapalumab can readily shut down IFN-γ signaling downstream of IL-18. Moreover, IL-18 itself can be targeted. The recombinant IL-18 binding protein (r-hIL-18BP), Tadekinig alfa, is currently under investigation in patients with MAS (NCT03113760) and refractory CRS/HLH (NCT05306080) and may serve as rescue agent in patients treated with IL-18 enforced T cell products. Similarly, the anti-IL-1β/IL-18 bi-specific antibody (MAS825), which is currently investigated in patients with monogenic IL-18-driven autoinflammatory diseases (NCT04641442), can deplete IL-18 and thus abrogate the dysregulated IL-18:IFN-γ axis (figure 1).

Next to such cytokine-targeted interventions, small-molecule inhibitors of Janus kinase (JAK) and signal transducer and activation of transcription (STAT) pathways (JAKis), that is, ruxolitinib and tofacitinib, are used for HLH and MAS treatment. Importantly, those inhibitors can disrupt cytokine-signaling pathways critically involved in hyperinflammation at once. JAK/STAT-inhibition affects IFN-γ signaling, but can also blunt type-1 interferon driven IL-18 expression⁴¹ and cytotoxic T cell hyperactivation.⁴² JAK/STAT-inhibition can furthermore affect downstream effects of several other pro-inflammatory cytokines with relevance in HLH and/or CRS such as IL-2, IL-6, IL-12 or GM-CSF. In mouse models of HLH, ruxolitinib reversed inflammatory responses and organ manifestations of HLH and enhanced survival, even more effectively than IFN-γ inhibition.⁴³ Ruxolitinib was successfully used as salvage and first-line therapy in both pediatric and adult patients with various manifestations of hyperinflammatory disease/HLH, including IEC-HS, standardized trials to follow.⁴⁴ However, the use of JAKis has been reported to increase the risk for opportunistic (viral) infections⁴⁵ and thus requires appropriate surveillance during treatment. As another drug with even broader effect compared with JAKis, rapamycin has recently been suggested as rescue therapy in cases of refractory sJIA-associated MAS, following the identification of a critical role of dysregulated mTOR-signaling in MAS pathology (figure 1).^{42 46}

One concern regarding the use of drugs with a broad/pan-inhibitory effect on cytokine signaling remains, in that their use may not only abrogate CAR T cell-induced toxicities but also compromise the overall T cell function and the therapeutic efficacy of CAR T cell treatment. Yet, their use may rather be considered as last resort treatment when life-saving measurements are clearly to prioritize over maintaining in vivo CAR T cell activity. Collectively, at present, there is an arsenal of drugs available which can help to control potential hyperinflammatory responses triggered by IL-18 and thus significantly mitigate potential risks of novel anticancer immunotherapies relying on this cytokine.