Indolent CD8+ CAR T-Cell Lymphoma of the Gastrointestinal Tract after Ciltacabtagene Autoleucel Therapy for Relapsed/Refractory Multiple Myeloma

Background:

T-cell lymphoma (TCL) derived from chimeric antigen receptor (CAR)-T cells have been reported as a rare event (2 cases) in multiple myeloma (MM) patients receiving ciltacabtagene autoleucel (cilta-cel; Harrison et al., ASH 2023; Ozdemirli et al., NEJM 2024), one of which was limited to the gastrointestinal tract. Here, we present the second reported case of enterotropic CAR+ TCL post cilta-cel, presenting as persistent diarrhea.

Methods:

Diagnostic workup for TCL included routine labs, serial EGD/colonoscopies (EGD/C) with biopsies with immunohistochemistry and FDG-PET scan. The presence of BCMA-CAR+ cells in tissue and blood was assessed by flow cytometry using fluorochrome labeled human BCMA/TNFRSF17 protein (CAR FACS). T-cell receptor (TCR) clonality was assessed by high-throughput sequencing (TCR-HTS). Mutational profiling of tissue and blood was performed by a clinically validated targeted next-generation sequencing, Stanford Tumor Actionable Mutation Panel for Hematopoietic and Lymphoid Neoplasms (Heme-STAMP).

Results:

A 50 y/o male with a history of kappa light chain MM received cilta-cel as 5th line of therapy. Upon evaluation for CAR-T, his free kappa light chain (FKLC) was 33.8 mg/dL, bone marrow plasma cells (BMPCs) >90%. After a bridging therapy with high-dose cyclophosphamide, the BMPCs decreased to 1-5% and FKLC to 11.5 mg/dL. After CAR-T infusion, he had grade 1 CRS on day (D)8 lasting 24 hours with no ICANS. On D59, he developed profuse watery diarrhea with negative work-up for infectious etiology and was treated with antibiotics with moderate improvement. On D86, he presented with 4-6 liters (L) stool/day. Biopsies of the stomach and duodenum showed crypt apoptotic activity. Of note, he did not have prior history of allogeneic transplant or inflammatory bowel disease (IBS). He received budesonide, corticosteroids, and infliximab (10 mg/kg) on D102/D111 with improvement to 4 formed stools/day. However, by D133, he had recurrent 4-6L diarrhea/day with 17 kg weight loss. Repeat EGD/C D137 revealed duodenal erosions and mild atrophy and edema in the proximal small intestine with dense T-cell infiltration in the lamina propria (predominately CD8+/TCRb+/TRBC1- cells positive for TIA1, granzyme, and perforin). H. pylori and celiac tests were negative. He received 2 more doses of infliximab. Repeat EGD/C on D151 showed villous blunting and foveolar metaplasia in duodenum and ileum with decreased density of lymphocytes. There was no evidence of chronic intestinal inflammation to suggest IBS. Due to persistent diarrhea (2L/day) he was started on IL-12/23 antibody ustekinumab on D178. However, due to minimal improvement, cyclosporine 5 mg/kg/day was started on D196. After attainment of target level (300-400 ng/mL), he began to have 4 formed stools/day D226, which is stable to date. A EGD/C D246 showed significant improvement of the duodenal architecture with reduced lymphocytes infiltration in the lamina propria.

FDG-PET on D210 showed no FDG-avid lesions. CAR FACS of D199 duodenum biopsy showed 35% CD8+CAR+ cells, 16% CD4+CAR+ cells of the total CD45+cells, respectively. Similar findings were observed from ileum (CD8+CAR+ 26%; CD4+CAR+ 13%). TCR-HTS showed 2 dominant nonproductive TCRg sequences, Vg2-Jg1/2 (40%) and Vg5-Jg1/2 (27%). In TCRb, 2 dominant sequences were identified, Db1-Jb2-1 (46%, nonproductive), Vb19-Jb2-3 (13%, productive). Retrospectively, these sequences were also dominant in D97 ileum and D144/D151 duodenum biopsy. Interestingly, these TCRb and TCRg sequences were also detected as dominant from D207 peripheral blood. Heme-STAMP on D144 duodenal biopsy revealed SH2B3 L390P at allele frequency of 26%, which was mapped to the signaling-inhibitory SH2 domain and the variant was predicted to be pathologic by 7 out of 7 variant effect prediction algorithms (Li et al., Sci Adv 2022). Genomic changes of SH2B3 are reported to contribute to oncogenesis in TCL (Bastidas Torres et al., Haematologica 2022). The clonal CAR+ T-cell proliferation within the GI tract was consistent with indolent CAR+ TCL .

Conclusions:

CAR+ TCL after CAR-T therapy is a rare entity. This is the second reported case of an enterotropic CAR+ TCL. A missense mutation in SH2B3 potentially contributed to the oncogenesis of TCL, but further investigation is needed to elucidate key mechanisms underlying the pathogenesis of this TCL.