Hh and NF-κB pathways are positively correlated in DLBCL, and Hh pathway contributes to NF-κB activation in DLBCL. (A) GLI1 expression levels in 145 DLBCL tumor biopsies. (B) High GLI1 expression was more frequent in DLBCL of ABC type than in GC type; in addition, low GLI1 expression was more frequent in cases of GC type than in those of ABC type. *P < .05. (C) A significant positive correlation between nuclear GLI1 and nuclear p-p65 (R = 0.3837, P = .0048) in a series of 54 DLBCL analyzed by immunohistochemistry (IHC). Representative IHC staining results for nuclear GLI1 and p-p65 in 2 DLBCL cases. Left (case 1), the tumor cells were negative for GLI1 and p-p65. Right (case 2), the tumor cells expressed both GLI1 and p-p65 in the nuclei. A subset of lymphoma-associated macrophages was also frequently and strongly positive for p-p65, which was confirmed by double immunohistochemical studies using p-p65 and the macrophage marker CD68 (supplemental Figure 1A). (D) Coculture of DLBCL cells (DOHH2 and HBL-1) with the stromal cell HS-5 in trans-well experiments resulted in increased p-p65. This effect was abrogated, at least in part, in the presence of the anti-Hh blocking antibody 5E1 (2.5 µg/mL). (E) DOHH2 and HBL1 cells treated with or without recombinant Shh (250 ng/mL) for 30 minutes or cyclopamine-KAAD (4.8 µM) for 1 hour were subjected to EMSA by using ³²P-labeled NF-κB or Oct-1 probes (Upper panel) as well as western blot to detect p-p65 (Lower panel). Dimethylsulfoxide (DMSO; vehicle for cyclopamine-KAAD) and phosphate-buffered saline (for recombinant Shh) were used as controls for drugs treatments.