ALL in escape room

Redirecting T-cell cytotoxicity through bispecific antibodies or chimeric antigen receptor (CAR) manipulation has become a key strategy in the treatment of B-cell precursor acute lymphoblastic leukemia (BCP-ALL). In this issue of Blood, Zhao et al decipher the mechanisms leading to primary resistance and relapse after treatment with CD19-targeted blinatumomab.¹  Through a comprehensive analysis of the leukemic cells and their environment, this study identified determinants of both primary resistance to and relapse after blinatumomab therapy.

Blinatumomab has been approved by the US Food and Drug Administration for the treatment of relapsed/refractory (R/R) and minimal residual disease (MRD)-positive BCP-ALL in children and adults. Following the pivotal phase 3 study in adults with R/R disease,²  2 recent and so far unpublished randomized studies showed a significant outcome improvement after blinatumomab-based consolidation in children, adolescents, and young adults in first high-risk relapse.^3,4  About half of the patients with R/R BCP-ALL are refractory to blinatumomab as single-agent salvage therapy, and the majority of them will relapse.

The efficacy of blinatumomab relies on the capacity to recruit patient immune effector cells potentially altered by the disease or previous therapies. T-cell expansion induced by blinatumomab is indeed predictive of response, but of limited interest as a biomarker. This expansion may be inhibited by high levels of regulatory T cells coactivated by blinatumomab.⁵  Resistance to blinatumomab may also be driven by the expression of immune checkpoint ligand programmed death-ligand 1 on leukemic cells.^6,7  In the current study, the authors used global and single-cell RNA-seq profiling in pretreatment samples to identify tumor- and immune environment-related determinants of resistance. Not surprisingly, increase in naive and central memory T cells was associated with a superior response rate, whereas an increase in exhausted T cells predicted a risk of failure. Moreover, nonresponder samples displayed a T-cell repertoire with clonal expansion and enriched in mucosal-associated invariant T (MAIT) cells, unconventional T cells involved in nonpeptidic antigen recognition.

The loss of CD19 antigen at relapse has been reported after both blinatumomab and CAR T-cell therapies. The landscape of mechanisms leading to CD19 negativity after CAR T has been investigated. These mechanisms include loss of heterozygosity, truncating mutations, and alternative splicing leading to the loss of the epitope encoded by exon 2.⁸  CD19-negative relapse after blinatumomab is observed in up to one-third of patients.⁸  Underlying mechanisms of resistance to blinatumomab were until now poorly explored. In the present study, Zhao et al provide a broad overview of these mechanisms, including truncating mutations, mutant allele-specific expression, decreased levels of CD19 RNA expression, and mutations in the CD81 gene encoding a tetraspanin responsible for CD19 maturation and transport to cell surface. None of these mutations were found before exposure to blinatumomab. Interestingly, the authors reported an association between a specific CD19 isoform associated with a partial deletion of exon 2 and both the risk of primary resistance and relapse.

So far, there are few studies that correlate ALL oncogenic events and response to blinatumomab. A study conducted in adults with R/R Philadelphia (Ph)⁺ ALL showed a response rate close to this reported in Ph⁻ ALL.⁹  Immature ALL presenting with KMT2A or ZNF384 rearrangements may be exposed to a higher risk of lineage switch and thus of escape through the loss of CD19 target. In the present study, patients with Ph-like ALL and CRLF2 overexpression were more likely to respond to blinatumomab. This observation is also supported by the RNA-seq profiling showing an enrichment of IL6-JAK-STAT pathway signature in responding patients, a hallmark of this subgroup of ALL. Conversely, patients with low hypodiploidy, including loss of chromosome 16 encoding the CD19 gene, were more likely to acquire a mutation on the remaining allele.

All these observations raise many questions for clinical practice. To increase the cohort size, the population of this study was enriched with relapsed cases, which in the absence of denominator limits correlation analysis with outcome data. More importantly, the cohort is mostly composed of R/R patients, which provides information that may be not extrapolated to patients treated with positive MRD.

1. Do we have well-recognized biomarkers to predict resistance to blinatumomab? The present and prior studies provide rational “intrinsic and extrinsic” candidates, including the presence of specific CD19 isoforms, or the percentage of peripheral regulatory T cells. These are more or less challenging to implement routinely and require prospective validation. Today, tumor burden, including extramedullary disease assessment, remains one of the most useful simple factors in practice, with again a lack of prospective validation and raising additional questions on the choice of debulking therapy.

2. Can we improve the efficacy of blinatumomab? In addition to promoting a favorable effector-to-target ratio, the most popular strategies supported by the present work involve modulations of immune environment. It is unclear if blinatumomab is more effective in the allogeneic setting, and exceptional cases of graft-versus-host disease have been reported. Trials assessing the role of blinatumomab early after allogeneic hematopoietic stem cell transplant or in combination with donor lymphocyte infusions are ongoing. The role of checkpoint inhibitors to overcome T-cell exhaustion are also being explored with different combinations and infusion timings. Targeting MAIT cells is an emerging field in cancer immunotherapy and questions about the potential implication of microbiota.¹⁰ 

3. Can we prevent the loss of the CD19 target? The absence of detectable pretreatment mutations and the presence of mutation hot spots highlighted by this study suggest that there is selective pressure of blinatumomab on CD19. The risk of occurrence of these stochastic events is supposed to increase with the number of leukemic cells, which again argues for tumor load control before initiating therapy. There is a warning signal that specific BCP-ALL subgroups may be more inclined to lineage switch or mutation acquisition, but the data are too scarce not to treat at this point. Additional strategies include the cotargeting of multiple antigens, including CD22, which may be achieved by inotuzumab ozogamicin or specific CAR T cells.

In summary, this article by Zhao et al is an important contribution to the understanding of escape strategies of BCP-ALL from CD19-targeted therapy. These observations support the prospective evaluation of new biomarkers, exploration of new mechanisms of resistance, and combination therapy trials to further improve patient outcome.