Siglec-6 CAR T: magic bullet for a moving target

In this issue of Blood, Jetani et al¹ explore models of CAR T-cell targeting of Sialic acid-binding Ig-like lectin-6 (Siglec-6) in acute myeloid leukemia (AML). Whereas CAR T cells are a useful new treatment for patients with refractory B-cell acute lymphoblastic leukemia (B-ALL), developing CAR T-cell therapies for AML has been hampered by a lack of suitable targets. Most obvious CAR targets for AML are also expressed on hematopoietic stem cells (HSCs) or myeloid cells. Given the propensity for CAR T cells to persist, prolonged myeloid aplasia would be expected if myeloid cells or HSCs were recognized. Importantly, Siglec-6 has no significant expression on HSCs or myeloid cells; consequently, Siglec-6 CAR T cells should spare normal hematopoiesis.

Targeting AML with CAR T cells poses additional challenges. AML is a heterogenous disorder arising from dysregulation at diverse points in myeloid differentiation. CAR T-cell targets should be broadly expressed on all AML subtypes and should target leukemia stem cells (LSCs), a cell population held responsible for relapse. However, LSCs may be phenotypically distinct from the bulk population and may differ among patients. Different LSC subpopulations may even be present in a single patient (see figure). Finally, potential AML targets should not be expressed on other critical nonmyeloid or nonhematopoietic cells.

The best explored AML target antigens are CD33² and CD123.³ These targets are not ideal, as they are expressed by normal myeloid cells and HSCs, respectively. With targets that do not spare normal hematopoiesis, 1 pragmatic strategy is to use CAR T cells to act as a “bridge” to allogeneic HSC transplant (allo-HSCT). In that case, conditioning chemotherapy should eradicate the CAR T cells and rescue the patient from aplasia. However, this strategy involves a lengthy period of myelosuppression and may be poorly tolerated. There are limited, early clinical data on CAR T-cell therapy in AML using such targets, but responses have been reported (reviewed in Mardiana and Gill ⁴).

More complex strategies have been developed. “Logic gating” of CAR signaling can restrict recognition to a pattern of antigen expression, thus improving specificity.⁵ For example, He et al⁶ described CAR T cells that mediated cytolysis only after joint recognition of CD13 and TIM-3. Given that AML blasts express both targets, but HSCs lack TIM-3, HSCs were relatively spared. Because of the difficulties in sparing normal hematopoiesis, a radical solution of concomitant administration of donor HSCs, genetically edited so as not to express cognate target, has been proposed.⁷ More recently, targeting CD70⁸ and TIM3⁹ has been suggested. These antigens are expressed by blasts and LSCs, but not on HSCs, although expression is found on NK cells and monocytes.

How does Siglec-6 fare against other targets and approaches? Jetani et al show that Siglec-6 is expressed in ∼60% of AML cases, expression is retained on LSCs, but there is little expression on HSCs. Further, expression within the hematopoietic system was restricted to memory B cell and basophil populations. Outside the hematopoietic system, no expression was identified other than the placenta. Furthermore, Jetani et al also demonstrate that Siglec-6 may be a useful target in B-Cell chronic lymphoblastic lymphoma (B-CLL). Siglecs are a large family of proteins that are thought to promote cell-cell interactions and regulate innate and adaptive immune systems through glycan recognition. It is worth noting that cancer immunotherapies targeting Siglecs are not new: CD22 (Siglec-2) and CD33 (Siglec-3) have been targeted with well-investigated immunoconjugates and CAR T cell therapies for ALL and AML, respectively.

The investigators next generated CARs based on a human monoclonal antibody (mAb) and demonstrated function in vitro, as well as in a range of AML models. Siglec-6 expression level is low but appears sufficient to direct CAR-mediated lysis against LSCs among other cells. HSCs were relatively spared after in vitro coincubation with Siglec-6 CAR T cells, although in vivo HSC engraftment studies were not described. One possible limitation of Siglec-6 is that proliferation and cytokine release was observed only when Siglec-6 CAR T cells were co-cultured with target cells expressing higher levels of target antigen.

In summary, Jetani et al expand the possibilities of CAR T-cell targeting in AML. Siglec-6 CAR T cell therapy could be useful for ∼60% of patients with AML and unlike many other targets, should spare normal hematopoiesis. It is hoped that clinical exploration of this and other strategies will bring CAR T cell therapy to patients with refractory AML.