Concurrent therapy of chronic lymphocytic leukemia and Philadelphia chromosome-positive acute lymphoblastic leukemia utilizing CD19-targeted CAR T-cells

Chimeric antigen receptors (CARs) are recombinant receptors composed of an antibody-derived extracellular single-chain variable fragment linked to a costimulatory signaling domain combined with the CD3-zeta chain of the T-cell receptor. T-cells modified to express anti-CD19 CARs, including 19–28z, incorporating the intracellular signaling domain of CD28, have demonstrated antitumor activity in patients with B-cell acute lymphoblastic leukemia (B-ALL) and chronic lymphocytic leukemia (CLL) [1–5]. Herein, we describe use of 19–28z CAR T-cells in a patient with concurrent CLL and Philadelphia chromosome-positive (Ph+) ALL.

The patient was diagnosed with CLL at age 47 following biopsy of enlarged cervical lymph nodes; initial cytogenetic and molecular data are unknown. After multiple short remissions following chemoimmunotherapy, he underwent double umbilical cord blood transplantation (dUCBT) seven years following the diagnosis of CLL (Table S1). Two years post-dUCBT, peripheral blood (PB) fluorescence-activated cell sorting (FACS) suggested relapsed CLL, and he received single-agent rituximab followed by maintenance rituximab for two years.

Fourteen years post-diagnosis of CLL (seven years post-dUCBT), he presented with weight loss, night sweats, cough, and pleuritic chest pain. Chest radiograph was consistent with pneumonia. CBC revealed WBC 29.3 k/μL with 95% mature lymphocytes, 2% blasts, hemoglobin 9.2 g/dL, and platelets 21 k/μL. Bone marrow (BM) biopsy demonstrated hypercellular marrow, with lymphoid blasts comprising 90% of cellularity. FACS and immunohistochemistry (IHC) confirmed markedly increased B-lymphoblasts and variable amounts of small, mature lambda-restricted B-lymphocytes, consistent with simultaneous involvement by B-ALL and CLL (Figure 1). Cytogenetic/FACS data are summarized in Table S2. FISH testing identified BCR-ABL1 rearrangement. He began therapy for B-ALL per CALGB 9511 [6]. Morphologic B-ALL (42% blasts by aspirate) persisted following course I (induction). Karyotype demonstrated 8/20 cells with abnormal recipient karyotype with BCR-ABL minor rearrangement and additional chromosomal abnormalities (46,XY,del(6) (q15q21), add (8)(p23), t(9;22)(q34;q11.2), −11), consistent with persistent Ph + ALL of recipient origin. BCR-ABL1 p190 fusion transcripts were detected. Upon beginning course II (intensification), numerous blasts expressing CD10, CD20, and TdT were noted in the CSF. He received six doses of intrathecal methotrexate with systemic chemotherapy. Dasatinib was initiated with course II, with dose increased on hematopoietic recovery.

Figure 1.

(A) BM aspirate demonstrating concurrent CLL and B-ALL and (B) PB smear demonstrating CLL cell predominance; Wright-Giemsa stain, 100x. (C) Hematoxylin and eosin stained serial sections (1) of BM along with IHC staining (2–6); 20x. Two distinct populations of malignant cells are present. One population (B-ALL) demonstrates positive co-staining for Tdt, Pax5, and CD34; the other (CLL) shows positive co-staining for CD5 and CD23 with dim CD20 expression (not-shown).

He was referred for investigational CAR T-cell therapy (NCT01044069). Following four weeks of systemic therapy per course II, BM biopsy showed B-ALL involving 30% of cellularity and small interstitial CD5⁺CD20⁺CD23⁺ lymphoid nodules (IHC), consistent with CLL. BCR-ABL1 p190 fusion transcripts (2%) were detectable. Autologous T-cells were collected, activated using Dynabeads^® CD3/CD28 CTS^™, transduced with a retroviral vector expressing 19–28z, expanded, and cryopreserved as previously described (Figure S1) [1,7]. Dasatinib monotherapy was continued during CAR T-cell manufacturing. Immediately prior to CAR T-cell administration (5 months post-diagnosis of B-ALL), BM biopsy demonstrated persistent CLL (40% of cellularity) and minimal residual Ph + ALL (BCR-ABL1 p190 transcripts, 0.02%).

He received cyclophosphamide 3 g/m² IV two days prior to CAR T-cell infusion, followed by 3.9 × 10⁶ 19–28z CAR T-cells/kg, with one-third of the total dose given the first day of infusion (day 0), and two-thirds administered the second (day +1). On day +1, he developed fevers and chills; broad spectrum antibiotics were initiated. He briefly developed fluid-responsive hypotension, but never required vasopressors/ICU care. Initial evaluation for infection was unrevealing, though parainfluenza virus type-3 was detected day +9. Fever subsided day +11. Mild lethargy was noted day +2, but no encephalopathy, seizures, or other neurologic adverse events occurred. He required neither tocilizumab nor corticosteroids and was discharged home day +13. Levels of several immunoregulatory cytokines peaked within 7 d of CAR T-cell administration, coinciding with maximal CAR T-cell expansion (Figure 2). Maximal detectable CAR T-cell persistence observed was 14 d in PB (by FACS/quantitative PCR) and 21 d in BM (by FACS).

Figure 2.

(A) Levels of immunoregulatory cytokines (CK) prior to and following conditioning chemotherapy and CAR T-cell infusion. (B) Vector copies/mL in PB post-CAR T-cell infusion. CD3⁺CAR⁺ % by FACS (flow) noted at specified time points as well. Cy: cyclophosphamide.

BM aspirate day +11 revealed no detectable BCR-ABL1 p190 transcripts. BM biopsy day +21 revealed no morphologic, immunophenotypic, or molecular evidence of B-ALL or CLL. He received intrathecal cytarabine alternating with methotrexate q 3–4 months for 2 years while continuing daily dasatinib and monthly intravenous immune globulin. CSF studies continued to show no evidence of B-ALL. BM aspirate, biopsy, and cytogenetic studies remained unremarkable 8.5 months post-CAR T-cell infusion, though FACS revealed a minute lambda-restricted population immunophenotypically similar to previously identified CLL, comprising 0.016% of total leukocytes. BM FACS 21 months post-CAR T-cell infusion again showed a persistent CLL population comprising 0.077% of total cells, also evident by PB FACS 25 months post-infusion (Figure S2). He has nevertheless required no further CLL-directed therapy. BM/PB studies have shown continued CR of Ph + ALL without detectable BCR-ABL1 p190 transcripts as of 43 months post-infusion.

Minimal residual disease (MRD) was additionally monitored by immunoglobulin heavy chain (IgH) deep sequencing (clonoSEQ, Adaptive Biotechnologies, Seattle). Prior to CAR T-cell infusion, BM clonoSEQ studies demonstrated two clonal IgH rearrangements comprising 72.7 and 10.5% of sequences. At 11 d post-CAR T-cell infusion, relative B-cell depletion was evident in BM, and neither sequence was detected. Thereafter, both IgH rearrangements became detectable at low levels in BM samples, with clonoSEQ studies estimating 2504 residual malignant cells per million at 12.9 months post-CAR T-cell infusion based on the dominant rearrangement (Figure S3). Whether both IgH rearrangements are associated with the CLL population, or whether the second, non-dominant IgH rearrangement detected at low levels is associated with the Ph + ALL population remains uncertain; BCR-ABL1 p190 transcript levels remain below the threshold of detection.

Concurrent development of CLL and lymphoblastic neoplasms, particularly Ph + B-ALL, is extremely rare [8]. In several cases, clonal relation between underlying CLL and subsequent Burkitt leukemia/lymphoma or B-lymphoblastic lymphoma has been suggested by new cytogenetic abnormalities superimposed on known karyotypic abnormalities present in the original CLL clone, expression of similar surface monoclonal immunoglobulin, or identical IgH rearrangements found in CLL cells, and the lymphoblastic neoplasm [9,10]. However, lymphoblastic neoplasms may arise as independent events in patients with preexisting CLL, without clonal association [8]. In this case, Ph + ALL and CLL populations were immunophenotypically distinct. Karyotype identified several additional chromosomal abnormalities in association with the Ph + ALL clone, but not deletion 13q or loss of 5′ IgH, as identified by FISH when morphologic evidence of CLL was present. Two separate clonal IgH rearrangements were additionally detected and monitored by clonoSEQ. Nonetheless, such studies do not entirely exclude lymphoblastic transformation of a CLL subclone.

Multiple factors may have contributed to long-term control of Ph + ALL with undetectable BCR-ABL1 p190 transcript levels, despite CLL progression, following 19–28z CAR T-cell administration. At CAR T-cell infusion, the patient had only minimal residual Ph +ALL and remained on dasatinib therapy, whereas morphologic evidence of CLL was present. The presence of a second IgH rearrangement detectable by clonoSEQ also suggests possible low-level persistence of a Ph + ALL clone in addition to the CLL clone, with ongoing dasatinib therapy suppressing Ph + ALL below the threshold of detection by BCR-ABL1 p190 transcript monitoring. Nonetheless, given the presence of residual Ph + ALL following consolidation chemotherapy in combination with dasatinib, long-term MRD-negative CR (by BCR-ABL1 p190 monitoring) with dasatinib alone would be unusual, suggesting 19–28z CAR T-cells deepened response. Maximal detectable persistence of 19–28z CAR T-cells was 21 d in BM by FACS (14 d by qPCR), suggesting long-term persistence was not required for durable control of Ph + ALL, though dasatinib activity may have contributed. An association between long-term CD19-targeted CAR T-cell persistence and ongoing CR among patients with R/R CLL has been observed in some but not all trials [1–3]. Each malignancy’s underlying ability to escape elimination by CAR T-cells, even within a single patient, may also have accounted for persistence of CLL despite long-term molecular remission of Ph + B-ALL. Additional in vitro studies to characterize the responsiveness of sorted malignant B-cell populations to CD19-targeted CAR T-cells, immune checkpoint ligand expression, and markers of CAR T-cell exhaustion at recovery could further support this hypothesis in future patients with multiple B-cell malignancies treated with this therapeutic modality, though would not fully recapitulate the endogenous tumor microenvironment.

B-cell malignancies differ considerably in the malignant cells’ interactions with the broader cellular environment. We and others have observed consistently high rates of MRD-negative CR among patients with relapsed/refractory B-ALL, including in patients with relapse following allogeneic hematopoietic cell transplantation. However, rates of durable MRD-negative CR have been lower among patients with R/R CLL [1–5]. Tumor immune resistance may account, in part, for differing responsiveness of B-ALL and CLL to second-generation CD19-targeted CAR T-cells in present clinical use. CLL cells exploit several mechanisms to evade the endogenous immune system, including induction of T-cell exhaustion and upregulation of inhibitory ligands impairing T-cell immunologic synapses [11,12]. Patients with CLL additionally exhibit increased numbers of regulatory T-cells and decreased T-cell proliferative capacity [13]. We and others are investigating whether engineering CD19-targeted T-cells to constitutively express additional co-stimulatory ligands or secrete pro-inflammatory cytokines may enhance CAR T-cell cytotoxicity and overcome inhibitory features of the CLL microenvironment in patients with R/R CLL [14,15]. Additional strategies under investigation include co-administration of antibody-based checkpoint blockade, or co-administration of ibrutinib therapy prior to T-cell collection and through CAR T-cell administration [13,16]. This report highlights the benefits of tumor-targeted cellular therapy and the differing responsiveness of these two B-cell malignancies despite the shared CD19 targeted surface antigen within the same patient, and therein illustrates an important ongoing challenge in the field.