- (1) Department of Laboratory Medicine and Pathology and.
- (2) Mayo Clinic, grid.66875.3a
- (3) University of Yamanashi, grid.267500.6
- (4) Aarhus University Hospital, grid.154185.c, Central Denmark Region
- (5) Department of Pathology and Laboratory Medicine, United Health Services Hospitals, Johnson City/Binghamton, NY.
- (6) China-Japan Friendship Hospital, grid.415954.8
- (7) University of California, Los Angeles, grid.19006.3e
- (8) Mayo Clinic, grid.417467.7
- (9) Odense University Hospital, grid.7143.1, Southern Denmark Region
- (10) Herlev Hospital, grid.411900.d, Capital Region
- (11) National University of Singapore, grid.4280.e
- (12) National University Cancer Institute, grid.440782.d
- (13) University of Iowa Hospitals and Clinics, grid.412584.e
- (14) University of Brescia, grid.7637.5
Anaplastic large cell lymphomas (ALCLs) are CD30-positive T-cell non-Hodgkin lymphomas broadly segregated into ALK-positive and ALK-negative types. Although ALK-positive ALCLs consistently bear rearrangements of the ALK tyrosine kinase gene, ALK-negative ALCLs are clinically and genetically heterogeneous. About 30% of ALK-negative ALCLs have rearrangements of DUSP22 and have excellent long-term outcomes with standard therapy. To better understand this group of tumors, we evaluated their molecular signature using gene expression profiling. DUSP22-rearranged ALCLs belonged to a distinct subset of ALCLs that lacked expression of genes associated with JAK-STAT3 signaling, a pathway contributing to growth in the majority of ALCLs. Reverse-phase protein array and immunohistochemical studies confirmed the lack of activated STAT3 in DUSP22-rearranged ALCLs. DUSP22-rearranged ALCLs also overexpressed immunogenic cancer-testis antigen (CTA) genes and showed marked DNA hypomethylation by reduced representation bisulfate sequencing and DNA methylation arrays. Pharmacologic DNA demethylation in ALCL cells recapitulated the overexpression of CTAs and other DUSP22 signature genes. In addition, DUSP22-rearranged ALCLs minimally expressed PD-L1 compared with other ALCLs, but showed high expression of the costimulatory gene CD58 and HLA class II. Taken together, these findings indicate that DUSP22 rearrangements define a molecularly distinct subgroup of ALCLs, and that immunogenic cues related to antigenicity, costimulatory molecule expression, and inactivity of the PD-1/PD-L1 immune checkpoint likely contribute to their favorable prognosis. More aggressive ALCLs might be pharmacologically reprogrammed to a DUSP22-like immunogenic molecular signature through the use of demethylating agents and/or immune checkpoint inhibitors.