Chronic Lymphocytic Leukemia (CLL) is a B cell malignancy that accounts for nearly 1/3rd of adult leukemia diagnoses in the Western world. Conventional chemo-immunotherapies initially control progression, but in the absence of curative options patients ultimately succumb to their disease. Chimeric Antigen Receptor (CAR) T cell therapy is potentially curative, but only 26% of CLL patients have a complete response. CLL-stimulated T cells have reduced effector functions and B-CLL cells themselves are believed to be immunosuppressive. Our work demonstrates that insufficient activation of CAR T cells by CLL cells mediates some of these effects and that the results are conserved between ROR1- and CD19-targeting CARs.
Results:
In this study we used an in vitro system to model the in vivo anti-tumor response in which CAR T cells serially engage with CLL cells. Multiple stimulations of CD19 or ROR1-targeting CAR T cells with primary CLL cells recapitulated many aspects of known T cell dysfunction including reduced proliferation, cytokine production, and activation. While the initial stimulation induced low level proliferation, subsequent stimulations failed to elicit additional effector functions. We further found that these functional defects were not permanent, and that CAR T cell function could be restored by switching to a stimulus with an aAPC (artificial Antigen Presenting Cell) control cell line. The aAPCs are well-characterized as potent stimulators of CAR T cell effector responses. Flow cytometry revealed that CLL-stimulated CAR T cells retained a non-activated, baseline differentiation profile, suggesting that CLL cells fail to stimulate CAR T cells rather than rendering them non-functional. One mechanism that could dampen activation is immune suppression. We assessed this at a high level by stimulating CAR T cells with CLL cells and aAPCs mixed at known ratios. However, even cultures containing 75% CLL cells stimulated proliferation and cytokine production.
Extensive immune-phenotyping revealed high level expression of the IL-2 Receptor on 90% (18/20) of the B-CLL cells tested. Since cytokine sinking via IL-2 receptor expression is a well-known mechanism of regulatory T cell suppression, we hypothesized that CLL cells similarly sink IL-2, blunting T cell activation. To test this, we supplemented IL-2 into CLL/CAR T cell co-cultures and showed that this rescued proliferation but only partially restored cytokine production. In contrast to our hypothesis, analysis of cytokine production by flow cytometry showed that CLL-stimulated CAR T cells did not produce IL-2 following a 6- or 12-hour stimulus, but TNFα was expressed after 12-hours. Similarly, CAR T cell degranulation, a prerequisite for target cell lysis was triggered after CLL recognition. These data again suggested that CLL cells insufficiently stimulate CAR T cell cytokine production, but also showed that cytolytic activity against CLL cells is intact.
We further proposed that CLL cells express insufficient levels of co-stimulatory and adhesion molecules to activate CAR T cells. Flow cytometry showed that most CLL cells expressed co-stimulatory and adhesion molecules at low levels; we hypothesized that up-regulating these molecules would enhance CAR T cell targeting of CLL cells. CLL cells were activated with CD40L and IL-4, which increased expression of CD54, CD58, CD80, and CD86. Stimulating CAR T cells with activated CLL cells enhanced CAR T cell proliferation and induced cell conjugate formation, indicating cell activation. Therefore, improving CLL stimulatory capacity can rescue T cell dysfunctions. To assess whether IL-2 addition and CD40 ligation were synergistic, we combined the two assays; however, we saw no additional improvement over IL-2 addition alone, suggesting that the two interventions may act upon the same pathway. Importantly, we also showed that rescue of CAR T cell function via IL-2 addition or CD40 ligation was not CAR-specific, as we observed the functional defects and subsequent rescue with both a ROR1-targeting CAR and the gold standard CD19-targeting CAR.
Conclusions:
Together, these data show that CAR T cell "defects" in CLL are actually insufficient activation, and improving the stimulatory capacity of CLL cells may enable better clinical responses. Further, this effect is not CAR-specific and these results may therefore be broadly applicable to multiple therapies for this disease.
Disclosures
Melenhorst: IASO Biotherapeutics: Membership on an entity's Board of Directors or advisory committees, Research Funding; Kite Pharma: Research Funding; Novartis: Other: Speaker, Research Funding; Johnson & Johnson: Consultancy, Other: Speaker; Simcere of America: Consultancy; Poseida Therapeutics: Consultancy.
LXR directly regulates glycosphingolipid synthesis and affects human CD4+ T cell function