Taking a BiTE Out of CAR-T Cell Efficacy

Multiple immunomodulatory therapies have emerged for B-cell acute lymphoblastic leukemia (B-ALL) treatment, changing the paradigm for relapsed or refractory disease. In a multicenter, retrospective cohort study, Myers et al¹ address an important question regarding the impact of sequential immunotherapy directed to the same antigenic target. Specifically, they examined how blinatumomab exposure before CD19-targeted chimeric antigen receptor T cells (CD19-CAR) affects response to cell therapy and survival.

THE TAKEAWAY

• In the article that accompanies this editorial, Myers et al¹ report on a multicenter, retrospective cohort study examining the impact of exposure to the CD19-direct bispecific T-cell engager, blinatumomab before CD19-targeted chimeric antigen receptor T cells (CD19-CAR) on outcomes. They found that blinatumomab nonresponse and high-disease burden at the time of CAR-T infusion correlated with worse outcomes, identifying important factors for prognosticating long-term outcomes after CD19-CAR.

They included 420 children and young adults treated with commercial product, tisagenlecleucel, or murine-based CD19-CAR T-cell therapy on six clinical trials at seven US centers. CAR constructs included two 4-1BB and one CD28 based CD19 constructs. Patients treated with blinatumomab were dichotomized for analyses as complete response (CR) or no CR. CR to CAR (CR-CAR) was defined as M1 marrow with < 5% lymphoblasts and no evidence of peripheral blasts or extramedullary disease 21-28 days after infusion.

CR-CAR was lowest for blinatumomab-exposed nonresponders (blina-no-CR; 64.5%), but comparable between blinatumomab-exposed responders (blina-CR; 92.9%) and blinatumomab-naive patients (blina-naive; 93.5%), P < .001. The cumulative incidence of relapse (CIR) was highest among blina-no-CR (CIR, 52.4%; 95% CI, 28.8 to 71.5), as compared with blina-CR (CIR, 26.2; 95% CI, 13.4 to 40.9%; P = .01) and blina-naive (CIR, 18.6%; 95% CI, 14.6 to 23.1; P = .0001). Although there was a trend toward lower CR rates in blina-naive as compared with blina-CR, the difference was not significant (P = .59).

Six-month event-free survival was poor after CD19-CAR in blina-no-CR patients (27.3%), as compared with blina-CR (66.9%, P < .001) and blina-naive (72.6%, P < .001). Six-month relapse-free survival (RFS) post-CAR was lower in blinatumomab-exposed patients 64.1% (95% CI, 50.4 to 74.9) versus blina-naive patients 81.1% (95% CI, 76.3 to 85.0), P = .02. However, inferior RFS was driven by the blina-no-CR group as RFS was not different in blina-CR and blina-naive groups.

Similar to prior studies,^2,3 high disease burden (≥ 5% marrow blasts) at the time of CAR infusion was associated with inferior outcomes. For blina-no-CR, high preinfusion disease burden further amplified poor survival.

CD19 expression by flow cytometry was evaluated before treatment with CAR-T and after blinatumomab if exposed. CD19 expression was stratified into four groups: positive (moderate-bright expression with > 90% blasts CD19+), dim (> 90% CD19+ blasts but less intensity than mature B cells), partial (population 50%-90% CD19+), and negative (> 50% of blasts CD19–). CD19 dim/partial expression was more common among the blinatumomab-exposed than blinatumomab-naive patients (13.3% v 6.5%, P = .06). Among patients with CD19-positive expression pre-CAR (n = 376), CR rates were lower in the blinatumomab-exposed, suggesting that additional factors other than antigen escape may influence CAR resistance. The investigators did not use more sensitive techniques to assess for rare CD19-negative subpopulations.

The majority of patients with CD19-dim disease pre-CAR experienced relapse (44.8%) or nonresponse (13.8%). Particularly poor outcomes were seen among the small cohort of blinatumomab-exposed patients who were CD19-dim pre-CAR (n = 9): RFS was < 25%. Among the CD19-positive cohort pre-CAR, CD19 expression at relapse did not substantially differ on the basis of prior blinatumomab exposure.

Myers et al¹ address important questions regarding the impact of blinatumomab exposure on post–CD19-CAR outcomes. These findings are timely as both are now US Food and Drug Administration–approved and European Medicines Agency-approved and used more frequently in ALL therapy. Indeed, multiple trials in adults and children are currently testing CD19-directed immunotherapy in the de novo setting.

A critical finding is that patients who do not respond to blinatumomab and proceed to CD19-CAR have worse outcomes. The authors reaffirm that higher disease burden also correlates with worse outcomes. One hypothesis has been that blinatumomab nonresponders and patients with high disease burden harbor difficult to detect CD19-negative clonal subpopulations that may portend worse outcomes. Myers et al¹ did find that CD19 dim/partial expression preinfusion was associated with lower event-free survival/RFS, and dim/partial expression was more frequent in blinatumomab-exposed patients. By contrast, a recent single-center study found that rare (< 1%) CD19-negative blasts or dim CD19 expression was not predictive of CD19-CAR response or disease recurrence, whereas blinatumomab exposure portended worse response and outcome, regardless of CD19 expression.⁴ This contrast highlights the need for more data and suggests that there may be complex mechanisms driving resistance to immunotherapy.

Importantly, Myers et al¹ found no significant difference in outcomes comparing blina-naive and blina-CR patients, suggesting that response to blinatumomab may be a surrogate for outcomes after CD19-CAR—yet prior blinatumomab exposure is not the main driver of resistance. As this was a retrospective study, these data did not capture patients who may have had CD19 escape after blinatumomab not referred to CAR-T. Accordingly, outcomes for the blinatumomab-exposed group could be skewed and prospective studies are needed.

A major question in the use of CAR-T therapy for patients with relapsed or refractory B-ALL is whether CAR-T is definitive therapy or best used as a bridge to hematopoietic stem-cell transplant (HSCT). Moreover, do prior exposure and/or response to blinatumomab affect the decision to use CAR-T as a bridge or stand-alone treatment? This study had the potential to address this important question but provided limited data on outcomes for the 146 patients who went on to HSCT post–CD19-CAR. Patients who did not respond to blinatumomab but subsequently responded to CD19-CAR had high CIR (> 50%), suggesting that in these patients they likely need additional therapy after CD19-CAR such as HSCT for cure. By contrast, those who responded to blinatumomab and subsequently responded to CAR-T had very good outcomes (CIR 26.2%). It is unclear whether these patients received HSCT post-CAR or CD19-CAR was used as definitive therapy. It is unfortunate these data are not included.

There were important differences in the blina-exposed versus blina-naive patients. Blina-exposed patients were more likely to harbor KMT2A-rearranged (KMT2Ar) B-ALL (14.3% v 6.7%, P = .04) and to have had prior HSCT (51.9% v 34.7%, P = .006). The presence of KMT2Ar is noteworthy as mechanisms of resistance to immunotherapy often differ on the basis of biology. KMT2Ar ALL has a high degree of lineage plasticity, and a key mechanism of failure is lineage switch.^5,6 By contrast, other types of B-ALL are more likely to relapse or develop resistance through CD19 downregulation or alternative splicing creating unrecognizable CD19 isoforms. Nevertheless, lineage switch was not seen more frequently in this blinatumomab-exposed cohort. Additionally, it has been hypothesized that CAR-T cell products generated in patients who had prior HSCT with persistent donor chimerism may be superior because of intrinsic differences in host T cells from patients with leukemia as compared with T cells from a healthy donor. The results were not stratified by prior HSCT, and there were many confounding factors, so the impact of post-transplant–derived T cells cannot be determined.

The potential impact of sequential immunotherapy with replicate antigen targeting is not specific to leukemia. In diffuse large B-cell lymphoma (DLBCL), there are several innovative therapies targeting CD19 including anti-CD19 monoclonal antibodies tafasitamab and loncastuximab tesirine and US Food and Drug Administration–approved CD19-CAR products axicabtagene, ciloleucel, and tisagenlecleucel. Does prior exposure to a monoclonal antibody or bispecific T-cell engager affect response and durability with subsequent CD19-directed therapy in DLBCL? Resistance to immunotherapy is likely more complex in lymphoma than in leukemia on the basis of differences in the tumor microenvironment. Nonetheless, these B-ALL data may suggest that patients with DLBCL who do not respond to one CD19-directed treatment may have worse outcomes with additional CD19-directed therapies.

Not all immunotherapies that target the same antigen are the same. The risk of antigen escape may differ with a monoclonal antibody as compared with a bispecific T-cell engager. The relative risk of sequential therapy on antigen escape, CR rates, and durability of response likely varies with immunotherapies targeting different antigens. Studies have clearly demonstrated that the mechanisms of escape or resistance are different with CD19- and CD22-directed therapies.⁷ When cytotoxic agents were introduced in the 1950s and small molecule inhibitors were introduced in the 1990s, it quickly became apparent that some phenomena are generalizable in cancer therapy—but many are not. Thus, caution should be taken before assuming any of the observations made in this study will translate to other immunotherapies or to other diseases.

With the emergence of new immunotherapies directed at the same targets, it is increasingly important to examine the risks and benefits of each treatment and the optimal sequence of administration. In the case of blinatumomab and CD19-CAR, blinatumomab response and high disease burden are clearly surrogates for nonresponse or high risk of relapse—but additional work is needed for understanding how to implement these findings into clinical practice.

AUTHOR CONTRIBUTIONS

Conception and design: All authors

Administrative support: David T. Teachey

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Taking a BiTE Out of CAR-T Cell Efficacy

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).