The CNS can be a safe space for CARs

In this issue of Blood, Frigault et al report that tisagenlecleucel, a CD19-specific chimeric antigen receptor (CAR) T-cell product, has a safety profile in a small series of patients with aggressive B-cell non-Hodgkin lymphomas (B-NHLs) and secondary central nervous system (CNS) involvement similar to that seen in patients without active CNS disease.¹  The Frigault et al study suggests that this product may benefit this previously excluded group of patients without an increased risk of neurotoxicity.

CARs combine the antigen-recognizing moiety of a monoclonal antibody with elements of the T-cell signaling machinery, most frequently a portion of the T-cell receptor–associated ζ chain and part of a costimulatory molecule such as CD28 or CD137 (4-1BB).²  T cells genetically engineered to express CARs targeting CD19 (CD19-CARTs) have demonstrated remarkable activity in patients with relapsed or chemotherapy refractory B-cell malignancies. In aggressive B-NHL, complete response rates ranged from 40% to 54% in the 2 largest published trials to date.^3,4  On the basis of this activity, 2 CD19-CART products, axicabtagene ciloleucel and tisagenlecleucel, have been approved by the US Food and Drug Administration (FDA).

Frequently complicating administration of CD19-CARTs, however, is the development of significant (black-box worthy) complications, including cytokine release syndrome (CRS) and neurotoxicity (recently named immune effector cell–associated neurotoxicity syndrome [ICANS] by a consensus panel).⁵  ICANS is a constellation of manifestations of CNS dysfunctions that include aphasia, altered levels of consciousness, impaired cognitive skills, motor weakness, seizures, and potentially cerebral edema, which has been fatal in patients who were treated in early-phase trials of other CD19-CART products.⁶ 

The underlying pathophysiology of ICANS has been difficult to elucidate. Expression of CD19 in the brain, which could lead to “on-target, off-tumor” effects, has not been clearly demonstrated. Recent human and animal data suggest that endothelial activation in the CNS and increased permeability of the blood-brain barrier after CAR T-cell treatment may expose pericytes to a high concentration of cytokines, triggering local inflammation.⁷  In contrast to CRS, however, interleukin-6 (IL-6) receptor antagonists, such as the monoclonal antibody tocilizumab, are not effective at reversing the manifestations of neurotoxicity. An IL-1 receptor antagonist, anakinra, may have therapeutic activity in this setting,^8,9  but experience with this agent is limited, and clinicians often must resort to using steroids, which have potential deleterious effects on CARTs and are not entirely effective at mitigating neurotoxicity.

In pivotal trials in B-NHL, the reported incidence of neurotoxicity ranges from 21% to 64% (12% to 28% grade 3 or 4).^3,4  However, the rate depends on the specific product and toxicity grading system used, which makes direct comparisons difficult. Nevertheless, patients with active CNS disease have been excluded from trials of CD19-CARTs for B-NHL because of concerns regarding the increased risk of neurotoxicity. Because CD19-CARTs are known to be able to traffic to the CNS,¹⁰  where they could potentially exert a therapeutic effect, these patients are possibly being denied a potentially beneficial treatment.

Frigault et al present a retrospective analysis of 8 patients with relapsed aggressive B-NHL with secondary CNS disease who were treated with lymphodepletion followed by tisagenlecleucel. Their data suggest that this approach has limited toxicity, because no patient required treatment of ICANS. In addition, these patients may benefit from CART therapy, since 4 of the 8 patients had partial responses. Although CART levels were not measured directly, an increase in circulating T cells and a rise in inflammatory markers were consistent with CART expansion, even in the absence of systemic disease. Two patients died with signs of increased intracranial pressure, but autopsy data, which the authors should be commended for obtaining, evidenced disease progression rather than CART-driven toxicity in the CNS.

Although this work describes a limited number of cases, it addresses a management issue for which there was only anecdotal evidence but that frequently arises in patients with relapsed or refractory aggressive B-NHL. These patients often have CNS disease and otherwise would be candidates for CD19-CART therapy. This series suggests that, with careful management (all patients received prophylactic anticonvulsants) and monitoring, at least some CD19-CARTs are a safe and potentially effective option for patients with secondary CNS B-NHL. Moreover, this report raises the possibility of using these cell products for primary CNS lymphoma, a diagnosis with far worse outcomes than its solely systemic counterparts that is not currently included in the indications for the FDA-approved products. Further studies in these settings are eagerly awaited.