CD19‐directed immunotherapy use in KMT2A‐rearranged acute leukemia: A case report and literature review of increased lymphoid to myeloid lineage switch

To the Editor: Acute lymphoblastic leukemia (ALL) bearing the KMT2A-rearrangement is a high-risk genomic subgroup, characterized by chemotherapy-resistance and a propensity for rapid relapse with reported 5-year overall survival of less than 20%.1 Unique in its involvement in early leukemogenesis, KMT2A gene rearrangement has been described to confer overlapping immunophenotypic features of both myeloblasts and lymphoblasts, hence previously known as the mixed-lineage leukemia (MLL) gene. This distinctive characteristic potentially drives lineage switch (LS) between lymphoid and myeloid leukemic cells during treatment, especially after highly lymphoid-targeted therapies. CD19-directed immunotherapies have revolutionized the treatment paradigm for relapsed/refractory ALL patients with the introduction of bispecific T-cell engaging antibody, blinatumomab, and novel chimeric antigen receptor T-cell (CAR-T) therapies. However, LS has been increasingly recognized as an emerging therapeutic challenge following these targeted treatments. Herein, we report the case of a Jehovah's Witness (JW) patient with relapsed KMT2A-rearranged ALL, who achieved an MRD (minimal/measurable residual disease)-negative, complete remission (CR) after one cycle of blinatumomab and subsequently presented with a LS to acute myeloid leukemia (AML). Moreover, we reviewed the literature and summarized all reported cases of ALL with LS following CD19-directed immunotherapy to help better characterize, identify the risk factors as well as potential genetic mechanisms driving this condition. A 60-year-old, JW female with Philadelphia chromosome-negative [Ph(−)], B-cell ALL was transferred to our facility following relapse of her disease. Initial cytogenetic and fluorescence in situ hybridization (FISH) studies revealed the KMT2A/AFF1 fusion [t(4;11)(q21;q23)]. No significant findings on next generation sequencing (NGS) were reported by the outside hospital. At diagnosis, the patient was initiated on a non-intensive induction regimen of hydroxyurea, prednisone, and vincristine. This was followed by POMP maintenance (vincristine, methotrexate, 6-mercaptopurine, and prednisone) after an MRD-positive CR was attained. The patient relapsed 7 months after initial diagnosis with a WBC exceeding 250 × 109/L for which she underwent leukapheresis. Available treatment options were limited as the patient declined transfusion of any blood products. The patient was commenced on blinatumomab salvage therapy and achieved an MRD-negative CR after her first cycle. No occurrences of severe neutropenia, thrombocytopenia, or anemia were encountered during her treatment course. Three weeks into her second cycle, her peripheral blood showed emergence of circulating blasts. Bone marrow cytologic study revealed AML with monocytic differentiation involving 80%–90% of total nucleated cells. Blasts expressing CD31, CD33, CD34 (small subset), CD43, CD68, and lysozyme were reported by immunohistochemistry; CD19 expression was not present on relapse. Cytogenetic studies revealed numerous chromosomal abnormalities in addition to the previously found KMT2A/AFF1 rearrangement, suggesting a LS with clonal evolution of the original leukemic clone. NGS revealed new mutations in TP53 and in RUNX1. Her pathology studies are available as Table S1. Treatment with venetoclax and azacitidine was initiated followed by venetoclax and decitabine to target her TP53 mutation. With the inability to support her with transfusions, despite supplementation with B12, folate, iron and growth factors, the patient was transitioned to comfort care and died of profound anemia while having active AML. Even though uncommon, conversion of leukemia lineage during treatment has been described in both AML and ALL patients. For ALL, this condition has most frequently been reported in infants and with the KMT2A-rearrangement.2, 3 Nevertheless, it remains a rare event with an estimated range between 0.6% to 6% at relapse.2, 4 Blinatumomab has demonstrated high rates of response with a favorable toxicity profile in the treatment of relapsed/refractory or MRD-positive ALL. However, key observations found among case reports of LS following CD19-directed immunotherapy raises concerns for KMT2A-rearranged patients (Table 1). A review of the past 8 years since the FDA's initial approval of blinatumomab identified 33 reported cases, 17 with blinatumomab and 16 with an anti-CD19 CAR-T cell therapy (Table 1).5-19 The median [IQR] number of prior CRs was 1 [1, 2] (among those reported) with time to LS occurring within 12 months in 90.9% of patients. Of significance, the KMT2A-rearrangement was implicated in 81.8% of cases—88.2% with blinatumomab and 75% with a CAR-T immunotherapy. A multicenter, retrospective study of 420 CAR-T patients by Lamble, et al. reported that the KMT2A-rearrangement was the predominant cytogenetic abnormality seen with LS, being present in 75% and 4.9% of LS and non-LS patients, respectively (p < .001).19 Age, gender, prior blinatumomab exposure, hematopoietic stem cell transplantation (HSCT), and CAR-T response were not significant predictors for LS in the investigators' analysis.19 Morbidity and mortality are high among patients with the KMT2A-rearrangement as intrinsic refractoriness to chemotherapy confers a dismal prognosis. CD19-directed immunotherapies have offered exciting treatment options for this high-risk subgroup, and in the case of blinatumomab, a bridge to allogeneic HSCT. However, high incidence of LS occurring with CD19-directed immunotherapies presents a unique predicament. LS to a myeloid or B-cell/mixed-phenotype leukemia rapidly occurred following blinatumomab initiation in the compiled KMT2A cases (n = 15) with 64.3% occurring at ≤1 month, 21.4% at >1 to 3 months, and 14.3% at >3 to 6 months (excluding the case of myeloid sarcoma at 8 months).10 This compared to anti-CD19 CAR-T therapy (n = 12) in which roughly a third of the cases were beyond 3 months (16.7% ≤1 month; 58.3% >1 to 6 months; 25% >6 months) (Table 1). Mortality was high following LS in patients treated with blinatumomab with CR achieved in only 38.6% (n = 5/13) of reported cases.5, 7, 8, 10-15 One patient experienced spontaneous disappearance of myeloid markers with cessation of blinatumomab but died shortly thereafter secondary to complications from ALL.9 The majority of CAR-T patients were from the study by Lamble et al. in which post-LS response rates were not reported; however, the investigators noted that there were no long-term survivors among their LS patients.19 Overall, 13% (n = 3/23) of KMT2A-rearranged patients with post-LS outcomes reported were alive at 1 year.5, 7, 8, 10-16, 19 A variety of mechanisms behind LS with ALL bearing the KMT2A-rearrangement after CD19-directed immunotherapies have been proposed.5-9, 11-14 Broadly, these hypotheses center around the ideas of clonal selection, bipotential leukemic progenitor cells, and cell reprograming.14 Selection against a dominant clone, allowing a minor resistant clone to expand has largely been discounted by other case reports.5-8, 12, 14 Similarly, there was no evidence of a secondary clone and many of the original clonal genetic characteristics persisted after LS with the patient case presented herein. A progenitor cell with both lymphoid and myeloid differentiation capabilities may play a role in LS especially since KMT2A-rearranged B-ALL has a pro-B phenotype with co-expression of myeloid antigens suggesting involvement of a common leukemic progenitor.14 However, this concept does not fully explain the often rapid and potentially higher incidence of LS in the context of blinatumomab exposure, specifically.20 Epigenetic activities and alterations of microenvironment modulators leading to cell reprogramming are another possible mechanism related to blinatumomab exposure. Inflammatory cytokines, notably IL-6, have been implicated as a driver of phenotypic changes.9, 11, 13 Loss of CD19 expression alone has not been demonstrated to result in LS, further supporting a more comprehensive cell reprogramming mechanism in the setting of the KMT2A-rearrangement.6, 8 A previous report has also suggested that a bulk population of lymphoid blasts underwent phenotypic switch simultaneously.5 While similar mechanisms for LS occurring after blinatumomab and CD19-directed CAR-T cell therapy have been proposed, the different temporal relationships observed between LS and exposure have not been elucidated. The pooled sample size is small and detailed characterization of this association is difficult at present; however, previously published case reports indicate that LS tends to occur more rapidly with blinatumomab (Table 1). The timeline for the patient presented fits with what has been previously described. It is possible that the different immune activation and responses after blinatumomab or CAR-T cell therapy may cause varying type or degrees of microenvironment modulator alterations. Lastly, the implication of sequential CD19-directed immunotherapy or consolidative HSCT on LS outcomes remains to be elucidated. Three patients were identified to have received blinatumomab prior to CAR-T therapy with early courses of LS occurring at 2-, 4-, and 5-months following cell therapy infusion.19 The interval from blinatumomab to CAR-T therapy was not reported. Among two other patients who proceeded to HSCT following blinatumomab, LS occurred between 2–4 months from blinatumomab initiation.13, 15 Significantly, in the study by Lamble et al., the investigators reported that eight of their 31 KMT2A patients who achieved CRs went on to receive consolidative HSCT and had no reported cases of LS with a median follow up of 1165 days.19 This compared to a 30.4% LS rate in the remaining 23 KMT2A-rearranged patients who did not receive consolidative HSCT, potentially demonstrating the importance of prompt HSCT to deeper and broader eradication of the leukemic clone in these patients. CD19-directed immunotherapies have emerged as a promising alternative to chemoimmunotherapy for refractory, high-risk ALL patients. However, patients with KMT2A-rearrangement present a unique challenge given the possibility of rapid occurrence of LS that ensues for many patients. An allogenic transplant may be planned in advance for patients with KMT2A-rearrangement, yet it should be noted that over 50% of reported LS cases in blinatumomab-treated patients occurred at less than 1 month. Further studies are warranted to identify the optimal therapeutic approach for this high-risk population. Until then, CD19-directed immunotherapies should be cautiously used in patients with the KMT2A-rearrangement, and consideration should be given to an urgent HSCT as soon as disease control has been achieved. Dr Alexandre Chan has consulting relationships with Eli Lilly and Company and Blueprint Medicines. Dr Susan O'Brien has consulting relationships with Amgen, Celgene, GlaxoSmithKline, AstraZeneca, Autolus, GlaxoSmithKline, Nova Research Company, Bristol Myers Squibb, DynaMed, Eli Lilly and Company, Janssen Oncology, Johnson and Johnson, Juno Therapeutics, MEI Pharma, Inc., Merck, Aptose Biosciences, Vaniam Group, AbbVie/Genentech, Sunesis Pharmaceuticals, Alexion Pharmaceuticals, Astellas Pharma, Gilead Sciences, Pharmacyclics, TG Therapeutics, Vida Ventures, and Pfizer and has received honoraria/research funding from Alliance, Janssen, Eisai, Amgen, Loxo Oncology, Inc., Mustang Bio, Inc., Nurix Therapeutics, Inc., Astellas Pharma, Aptose Biosciences, Beigene, Ltd., Caribou Biosciences, Inc., Acerta Pharma, Regeneron, Gilead Sciences, Pfizer, AbbVie, Alexion Pharmaceuticals, TG Therapeutics, Pharmacyclics, and Kite, a Gilead company. Dr Deepa Jeyakumar has received research funding through Pfizer and Jazz Pharmaceuticals. Dr Pamela Becker reports institutional research support from Glycomimetics, Notable Labs, and Pfizer, and advisory role at Accordant Health Services (CVS Caremark). Dr Piyanuch Kongtim has consulting relationships with CareDx. The remaining authors report no potential conflicts of interest. All data generated or analyzed during this study are included in the published article. Table S1 Pathology studies of bone marrow aspirates at initial leukemia diagnosis and post-lineage switch. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.