Anti-CD1a CAR T cells to selectively target T-ALL

In this issue of Blood, Sánchez-Martínez and colleagues present a novel approach to the treatment of T-acute lymphoblastic leukemia (T-ALL) using chimeric antigen receptor (CAR) T cells redirected against CD1a¹  in in vitro and xenograft models.

Recently, considerable progress has been made in the treatment of relapsed/refractory B-acute lymphoblastic leukemia (B-ALL) by application of antigen-directed immunotherapies such as blinatumomab, inotuzumab, and anti-CD19 CAR T cells.²  Of these new approaches, perhaps the most promising is anti-CD19 CAR T cells, which can induce long-lasting remissions.³ 

Given the similarities between B- and T-ALL, development of CAR T-cell therapy against T-ALL seems a rational approach. However, equivalent progress has not been made,⁴  mostly because CAR T-cell targeting of T-ALL is more difficult than that of B-ALL. CAR T-cell therapies against B-ALL target pan B-cell antigens. This strategy is feasible because the concomitant profound and prolonged B-cell aplasia is tolerable. However, the T-cell aplasia following CAR targeting of pan T-cell antigens would be prohibitively immunosuppressive. Furthermore, expression of the target antigen on the CAR T cells themselves may cause so-called “fratricide,” where CAR T cells kill each other.

A recent approach has been described where 1 of 2 alleles at the T-cell receptor (TCR) β constant region is targeted, allowing preservation of approximately half of the normal T-cell compartment. However, this approach is mostly applicable to mature T-cell malignancies, because only a minority of cases of T-ALL express surface TCR.⁵  An alternative approach is to generate CAR T cells targeting a pan T-cell antigen, from T cells in which expression of the pan T-cell antigen has been disrupted. Such disruption is typically achieved by using genome editing, and target antigens proposed for such a strategy include CD5⁶  and CD7.^7,8  This approach prevents fratricide but will likely lead to T-cell aplasia, which would require rescue by allogeneic hematopoietic stem cell transplant. A simpler strategy would be possible if an antigen selectively expressed by T-ALL blasts but not by normal T cells was targeted.

In the present study, Sánchez-Martínez et al propose CD1a as a selective T-ALL target. CD1a is a transmembrane glycoprotein that presents self- or bacterial-derived lipids to specialized T cells. It is present on ∼40% of cases of T-ALL, where its expression defines cortical T-ALL. CD1a expression on normal tissues is limited to a subset of skin-resident dendritic cells (Langerhans cells, LC) and a developmentally transient thymocyte population: crucially, it is not present on mature T cells. Sánchez-Martínez et al demonstrate that anti-CD1a CAR T cells could be generated without the use of genome editing, were not prone to fratricide or exhaustion, were able to efficiently lyse cortical T-ALL blasts in vitro and in multiple murine models, but spared normal T cells. Importantly, when anti-CD1a CAR T cells were incubated with fetal thymocytes, the majority of cells were preserved, indicating this therapy might not induce thymic ablation.

There are however limitations to CD1a as a target for T-ALL. First, only a minority of cases of T-ALL will express this antigen. CD1a has been associated with relatively good prognosis⁴ ; thus, the proportion of patients with CD1a⁺ relapsed and refractory disease will be small. Furthermore, in those cases that do express CD1a, blast expression of the target may not be uniform, leading to the possibility of selection of preexisting CD1a negative blasts. Another mechanism for escape might be downregulation of target, which does not seem critical for blast survival. Furthermore, the clinical consequences of depletion of CD1a-expressing thymocytes and LC are yet to be completely elucidated. LCs are the predominant antigen-presenting cells in the skin and are also present at other mucosal surfaces: their absence could lead to localized immunodeficiency, autoimmunity, or keratosis.

The work by Sánchez-Martínez et al forms part of a growing literature describing CAR T-cell strategies for T-ALL. In addition to the approach where fratricide is prevented by genome editing, we anticipate continuing efforts to develop alternative targeting approaches that avoid T-cell aplasia. These may include CARs against combinatorial “logic-gated” antigen pairs to target pan T-cell antigens without inducing T-cell aplasia. Alternatively, multiplexed targeting of selective antigens such as CD1a with overlapping expression profiles may prevent antigen-negative escape and broaden the pool of potential recipients. This study by Sánchez-Martínez et al is an important first step in practicable CAR T-cell therapy for T-ALL and offers hope that the considerable success of CAR T cells in B-ALL can soon be extended to patients with T-ALL.