Targeting Ovarian Cancer with IL-2 Cytokine/Antibody Complexes: A Summary and Recent Advances

Yilun Deng1, Ryan M Reyes2,3, Chenghao Zhang1,4, José Conejo-Garcia5, Tyler J. Curiel1,2,3,6* 1Department of Medicine, University of Texas Health San Antonio, San Antonio, TX, USA 2South Texas Medical Scientist Training Program, University of Texas Health San Antonio, San Antonio, TX, USA 3Mays Family Cancer Center, University of Texas Health San Antonio, San Antonio, TX, USA 4Xiangya Medical School, Central South University, Changsha, Hunan, China 5H. Lee Moffitt Cancer Research Institute, Department of Immunology, Tampa, FL, USA 6Graduate School of Biomedical Sciences, University of Texas Health San Antonio, San Antonio, TX, USA


Cancer Model
Recently, we reported beneficial treatment effects of IL-2c in a mouse ovarian cancer model [32] (major findings summarized in Figure 2), a significant advance in ovarian cancer (OC) immunotherapy for several reasons. OC is the deadliest gynecological cancer, leading to an estimated 13,700 deaths in the United States in 2021 [33]. It should respond to immunotherapies since intratumoral T cell infiltration predicts improved OC patient survival [34] which we showed is reduced by intratumoral Tregs [35]. However, decades of OC immunotherapy trials done by outstanding groups have been largely disappointing [36], which could be in part due to OC's relatively low mutation burden compared to other tumors [36].
We found IL-2c is highly efficacious in the aggressive and poorly-responsive ID8agg mouse ovarian cancer line that we developed, including being refractory to immune checkpoint blockade (ICB) with α-programmed death ligand 1 (PD-L1). Consistent with known IL-2c effects, IL-2c promoted the activation of effector CD8 + T cells and CD4 + non-Treg cells and increased production of antitumor cytokines including IFNγ and TNFα by these cells in ID8agg challenge. Surprisingly however, IL-2c also expanded the number of Tregs thus reducing the CD8 + T cell/Treg ratio specially in ascites, an effect not previously reported with this approach, which could reflect the OC TME or OC cells themselves. Further investigation revealed that IL-2c treatment impaired Treg suppressive function in a TMEspecific manner as Tregs from tumor draining lymph nodes (TDLN) suppressed normally.
Although the half-life of IL-2c is around 24 hours in vivo [11], we observed significant tumor-inhibitory effects as long as three weeks after the last dose of IL-2c. These data suggest induction of immune memory, evident by observed increases in central and effector memory CD8 + T cells after IL-2c treatment. Because IL-2c can increase TILs and tumors with pre-existing T-cell infiltration respond well to ICB [37], we explored combination of IL-2c and αPD-L1 therapy and found a synergistic effect of IL-2c plus αPD-L1 that was superior to either agent alone, in both (no t)ID8agg OC and B16 melanoma model. Moreover, mice that recovered from primary ID8agg challenge after combination therapy were resistant to subsequent tumor challenge, confirming induction of protective immune memory. Thus, IL-2c can work through simultaneous targeting of distinct immune pathways with effects that include: 1) activating effector T cells and promoting their antitumor cytokine production, 2) inhibiting immune-suppressive Tregs, and 3) inducing durable, protective immune memory.
In recent work, we performed additional exploration into the specific immune population(s) responsible for IL-2c efficacy in ID8agg tumors. CD8 + T cells are the primary target populations of IL-2c and are required for successful IL-2c treatment in several tumor models [9,14]. However, we found IL-2c was efficacious in ID8agg tumors following antibodymediated CD8 + T cell depletion in vivo depletion. We tested two depletion timepoints, during the course of IL-2c treatment starting in week 2 or a week post the last dose of IL-2c (week 3), both of which significantly depleted CD8 + TILs for the duration of the experiment ( Figure 3B) and did not affect IL-2c efficacy ( Figure 3A). This result was striking because despite IL-2c mediated activation of CD8 + TILs in other models (Table 1), they were dispensable for IL-2c efficacy here. We also recently published that in orthotopic MB49 mouse bladder cancer, CD8 T + cells were not required for IL-2c efficacy, which instead relied on δ T cells [31]. However, IL-2c treated ID8agg tumors effectively in γδ T cell-deficient mice [31], consistent with TME-specific IL-2c effects.

IL-2c Induces TME Treg Fragility Independent of CD8+ and δ T cells
Tregs are a major immune-suppressive population in the TME and are associated with worse outcomes in OC patients [35]. Tregs also play a crucial role in the development of ID8agg tumors as we previously showed that depleting Tregs completely abolished ID8agg tumor progression [32]. Strikingly, fragile Tregs were first described by the Vignali group and are characterized by maintained FoxP3 expression with loss of suppressive function and decreased production of effector molecules including granzyme B, versus an unstable Treg phenotype where Tregs lose their FoxP3 expression [38]. Fragile Tregs also express transcription factor T-bet and produce cytokines that are uncharacteristic of normal Tregs [39], including IFNγ. Treg fragility can also contribute to ICB treatment efficacy [38,39] which could be another mechanism of IL-2c synergy with αPD-L1. We found that Tregs displayed a fragile phenotype induced by IL-2c, which helps explain IL-2c efficacy despite increasing Treg numbers in the OC TME. We performed t-distributed stochastic neighbor embedding (tSNE) analysis of Tregs in the OC TME and TDLN and found that Tregs in TDLN treated with control or IL-2c have a similar immune phenotype, whereas IL-2ctreated Tregs in the ascites acquired the fragile phenotype ( Figure 4). By contrast, IL-2c did not induce Treg fragility in the bladder in MB49 bladder cancer, consistent with TMEspecific fragility induction.
We further explored mechanisms for IL-2c-induced Treg fragility. We observed that IL-2cinduced fragility can be detected three weeks after the final dose of IL-2c, but not one week after the final dose [32], and is likely indirect from CD25 signals. Ex vivo treatment of Tregs with IL-2c does not fully recapitulate the in vivo fragile phenotype [32], suggesting IL-2c cooperates with another factor or population in the OC TME to induce Treg fragility, or could work during Treg differentiation, which could account for the long time needed for in vivo fragility induction. IFNγ is a potent Treg fragility inducer [38] and IL-2c promotes IFNγ production in effector T cells. We assessed major IFNγ-producing population effects on Treg fragility induction, including CD8 + T cells and γδ T cells. Antibody-mediated CD8 + T cell depletion did not rescue the Treg fragility phenotype in ascites ( Figure 5A), and fragile Tregs with increased T-bet, and IFNγ expression were also present in ascites of γδ T cell deficient mice challenged with ID8agg tumors ( Figure 5B) excluding δ T cells and making CD8 + T cells unlikely mediators in IL-2c induced Treg fragility. Future studies will test NK cell, innate lymphoid cell and fragile Treg interferon-γ in IL-2c induced Treg fragility and assess other tumors and TME.

Conclusions and Future Directions
Our published studies [31,32] and new data shown here highlight several important mechanisms that could be unique to the OC TME and serve as important contrasts to other studies using the same IL-2c or similar CD122-biased modified IL-2. First, conventional CD8 + T cells are not required for the treatment efficacy of IL-2c in ID8agg OC, despite IL-2c-mediated promotion of IL-2 signaling and antitumor cytokines production in intratumoral CD8 + T cells. In this specific TME, IL-2c preferentially activates other immune populations that play a more crucial role than CD8 + T cells in treatment efficacy. Absence of OC TME CD8 + T cells is not seen and thus agents such as αPD-L1 that do activate OC TME CD8+ T cells can synergize with IL-2c as we reported. Other IL-2c activated populations including NK cells or other members of the innate immunity could be preferentially activated over CD8 + T cells or insufficient CD8 + T cells could induce compensatory activation of other cell populations. Understanding the preferential and distinct IL-2 signaling and activation of varied immune populations in specific TMEs and/or tumors is essential for comparing differences in IL-2 targeting agent mechanisms between various tumors and anatomic locations, and to define mechanisms for agents that can cooperate clinically with directed IL-2 agents, which to date is little studied. Second, IL-2c inhibits intratumoral Treg function by inducing a fragile phenotype at least in some tumors and environments. CD122-selective IL-2 agents are expected to avoid activation and proliferation of Tregs and most previous studies in other models observed an increased CD8 + T cell/Treg ratio [8,14,22,40] and sometimes even Treg depletion [19]. However, we found increased number of Tregs and fragility in ascites after IL-2c treatment. Studies are warranted to identify fragility mediators, define in what other cancer models or human cancers fragile Tregs can be induced and to understand which agents best combine with this effect for treatment efficacy.
Currently, many modified IL-2 agents are in varied stages of preclinical and clinical development [23] (and Table 1), yet a detailed understanding of the immune landscape during treatment is lacking in most cases. Our prior and ongoing investigations have uncovered unexpected yet useful immune consequences of IL-2c signaling in the OC TME, stressing the importance of understanding the biology and immunology of distinct IL-2 agents to advance our understanding of IL-2 and IL-2 receptor biology, and optimal therapeutic uses. We also shed light on OC-specific TME phenotypes that potentially help explain its unresponsiveness to current immunotherapies, which should facilitate further design of targeted immunotherapies for OC patients.    CD8 + conventional T cells are not required for the treatment efficacy of IL-2c. A. ID8agg tumors measured by luciferase as described [30].  tSNE analysis of ascites Tregs in ascites and tumor draining lymph nodes from isotype or IL-2c treated mice challenged with ID8agg. Flow cytometry analysis performed three weeks post last IL-2c as described [30]. Density plots show highly different Treg phenotypes among ascites and tumor draining lymph nodes from isotype and IL-2c treated mice challenged with ID8agg. IL-2c treated ascites Tregs (red) exhibited increased expression of fragility markers including T-bet and IFNγ compared to isotype treated ascites Tregs (orange), among other changes. Tregs from tumor draining lymph nodes of two groups of mice are very similar to each other (blue).