Anti-CD22 CAR-T (Xinqiao Hospital ARMY Medical University)

To the Editor: Chimeric antigen receptor T-cell (CAR-T) therapy has revolutionized the care of patients with relapsed or refractory (R/R) large B-cell lymphoma (LBCL). Tisagenlecleucel and axicabtagene ciloleucel are two anti-CD19 CAR-T therapies. They are approved by the United States Food and Drug Administration for the treatment of R/R LBCL, after two or more lines of systemic therapy. Despite unprecedented efficacy of CAR-T therapy in this multiple R/R setting with overall responses up to 83% and 58% complete responses,1 relapses post CAR-T therapy remain challenging to treat. In patients with B-acute lymphoblastic leukemia (ALL), resistance to CAR-T therapy is thought to be due in part to emergence of CD19 negative escape variants2; however, this mechanism of relapse is not well-established after CAR-T therapy for r/r LBCL. Herein, we report two patients with R/R LBCL who experienced a CD19-negative disease relapse following anti-CD19 CAR-T therapy. Two patients with CD19-negative disease relapse were identified from the University of Maryland Cellular Therapy database. The study and analysis was approved by the local institutional review board. Patient 1: A previously healthy 60-year-old woman was diagnosed with LBCL, treated with R-CHOP, and subsequently achieved durable complete remission (CR). Fourteen years later, she developed new adenopathy, and a lymph node biopsy showed small lymphocytic lymphoma (SLL) without peripheral blood lymphocytosis. Two years later, her lymphadenopathy worsened, and a positron emission tomography/computed tomography (PET/CT) showed diffuse, FDG-avid lymphadenopathy above and below the diaphragm. A lymph node biopsy showed high-grade B-cell lymphoma, with CD19 antigen expression, lambda light-chain restriction, and c-MYC and BCL6 rearrangement by fluorescence in situ hybridization (FISH). She received two cycles of R-ICE (Rituximab-Ifosfamide/Carboplatin/Etoposide) salvage therapy and attained a partial response (PR). However, prior to proceeding with an allogeneic hematopoietic stem cell transplant (allo-HSCT), she was found to have disease progression. Further salvage treatments with onalespib (HSP90 inhibitor) based clinical trial failed to produce any response. Subsequently, she received axicabtagene ciloleucel anti-CD19 CAR-T therapy following lymphodepletion with fludarabine and cyclophosphamide. Her clinical course was notable for grade 3 cytokine release syndrome (CRS) without neurotoxicity, which was treated per guidelines. One month post-CAR-T therapy (D + 30), PET/CT demonstrated resolution of all prior FDG-avid lesions (cervical, axillary, mediastinal lymphadenopathy). However, there was interval development of several hypermetabolic nodes in the abdomen and pelvis. Further progression was noted on a repeat PET/CT scan obtained 2 weeks later, and newly demonstrated, FDG-avid subcarinal node biopsy confirmed relapse of the LBCL. Flow-cytometry identified a clonal population of lambda light-chain restricted large lymphoma cells with loss of CD19 antigen expression. This was further confirmed by immunohistochemistry (IHC) (Figure 1). Moreover, program cell death ligand-1 (PD-L1) using the PD-L1 IHC 22C3 pharmDx assay, showed strong membranous staining on most of the large, atypical cells. Based on this information, the patient was treated with ibrutinib, pembrolizumab, and rituximab. Two months later, a PET/CT demonstrated PR which was maintained until 4 months total at which point the patient developed progressive disease. Patient 2: A 56-year-old man was diagnosed with follicular lymphoma (grade 2) on a colonoscopic biopsy performed to evaluate hematochezia. Imaging demonstrated widespread lymphadenopathy, and bone marrow biopsy and aspiration showed involvement by follicular lymphoma consistent with stage IV disease. He obtained a CR after treatment with bendamustine and rituximab followed by rituximab maintenance. Two years later, the patient developed weight loss and abdominal pain; imaging demonstrated a new abdominal mass. A biopsy revealed GCB-subtype, double-hit LBCL with c-MYC and BCL2 gene rearrangements by FISH, and B-cells expressed CD19. The patient underwent three cycles of salvage R-ICE therapy with achievement of CR. He declined an allo-HSCT and opted for rituximab maintenance. Seven months after the last CR, he developed new constitutional symptoms, and a PET/CT scan suggested relapsed aggressive B-cell lymphoma, for which he received salvage R-CHOP therapy, but showed no disease response. Given the patient's refractory disease, he received axicabtagene ciloleucel anti-CD19 CAR-T therapy, with fludarabine and cyclophosphamide lymphodepletion. His hospital course was uncomplicated, without development of significant CRS or neurologic toxicity. The D + 30 PET/CT scan imaging showed PR. Two weeks later, he developed abdominal distention from ascites, elevated lactate dehydrogenase and liver enzymes. A CT abdomen showed ascites, worsening of abdominal disease, and new omental deposits. Ascitic fluid cytology demonstrated involvement by his aggressive LBCL. While flow cytometry is the most sensitive modality to assess surface antigen expression, the patient's lymphoma cells did not survive processing, likely due to their large size. The CD19 IHC demonstrated dim to negative cytoplasmic staining in the neoplastic cells and without discernible CD19 surface expression.(Figure 1 ) This may either be secondary to loss of CD19 expression or internalization of the CD19 antigen. The patient was transitioned to hospice and succumbed to his disease. Here, we report two patients who received anti-CD19 directed CAR-T therapy for multiply R/R DLBCL, whose disease relapsed rapidly after loss of CD19 antigen and/or immune escape. This mechanism of resistance is well established in patients with B-ALL, whose disease relapsed with CD19 antigen escape following anti-CD19 CAR-T therapy,2 but limited published data exists in LBCL. Long-term follow-up of the ZUMA-1 study wherein patients received CD3ζ/CD28-based anti-CD19 CAR-T therapy showed that of 11 biopsy-evaluable patients with progressive disease, three showed loss of CD19 antigen expression following treatment with anti-CD19 CAR-T therapy.3 Furthermore, in the JULIET clinical trial with CD3ζ/4-1BB-based anti-CD19 CAR-T therapy, one out of five biopsy-evaluable patients with progressive LBCL had loss of CD19 expression.4 At our institution, 11 patients with relapsed disease after CAR-T therapy underwent biopsy. At the time of this analysis, 34 patients underwent CAR-T therapy for standard of care indication for DLBCL. Only two patients had evidence of CD19 antigen expression loss. This immune antigen escape mechanism has also been described in a patient with LBCL, who had CD22 loss after anti-CD22 CAR-T therapy.5 Thus, it is plausible that antigen-specific cellular therapy promotes clonal evolution permitting tumor immune evasion by selecting for down-regulated expression of the targeted antigen. To our knowledge, this is the first descriptive case series with detailed histopathologic analysis of CD19 immune escape following CD19 CAR-T therapy for LBCL. In patients with B-ALL, mechanisms of relapsed disease after CD19 targeted therapy include CD19-negative immune escape,2 as selective pressure through targeted therapy leads to the selection of pre-existing, or evolution of, antigen-negative clones. In order to overcome this resistance mechanism, further tailoring of the CAR construct is necessary in an attempt to produce enhanced antitumor activity and prevent expansion of individual antigen negative-clones. Phase I data from a bispecific CAR-T construct targeting CD19 and CD20, recently demonstrated comparable results with an overall response rate of 82% and CR in 55% of patients.6 Most importantly, however, was that none of the relapses resulted in down-regulation of target antigens. Real world analysis of relapses post CAR-T cell therapy will provide incidence of CD19 antigen escape. Pathologic examination of relapsed disease following CAR-T cell therapy will help us understand important mechanisms of relapse and develop more effective treatment strategies. We appreciate the outstanding work of the Cell Processing and Apheresis laboratories of the University of Maryland Greenebaum Comprehensive Cancer Center as well as the excellent care by inpatient and outpatient nurses and Medical Intensive Care Unit (MICU) staff. EH is on the Kite Pharma Speakers Bureau for Axicabtagene ciloleucel. KR and SD served as advisor for Kite Pharma, a Gilead Company. None of the other authors have any relevant conflicts of interest to disclose. APR and SD identified the described cases. AB wrote the manuscript with assistance from FEC and SD. ZS and RK provided interpretation and delivery of the images. All authors provided critical feedback and edited the manuscript.