Repurposing NAMPT Inhibitors for Germinal Center B Cell-Like Diffuse Large B-Cell Lymphoma

Abstract

Diffuse large B-cell lymphoma (DLBCL) includes the activated B cell-like (ABC) and germinal center B cell-like (GCB) subtypes, which differ in cell of origin, genetics, and clinical response. By screening the subtype-specific activity of 211 drugs approved or in active clinical development for other diseases, we identified inhibitors of nicotinamide phosphoribosyl transferase (NAMPTi) as active in a subset of GCB-DLBCL in vitro and in vivo. We validated three chemically distinct NAMPTis for their on-target activity based on biochemical and genetic rescue approaches and found the ratio between NAMPT and PARP1 RNA levels was predictive of NAMPTi sensitivity across DLBCL subtypes. Notably, the NAMPT:PARP1 transcript ratio predicts higher antitumor activity in BCL2-translocated GCB-DLBCL. Accordingly, pharmacologic and genetic inhibition of BCL2 was potently synergistic with NAMPT blockade. These data support the inhibition of NAMPT as a therapeutically relevant strategy for BCL2-translocated DLBCLs. Significance: Targeted therapies have emerged for the ABC subtype of DLBCL, but not for the GCB subtype, despite the evidence of a significant subset of high-risk cases. We identify a drug that specifically targets a subset of GCB-DLBCL and provide preclinical evidence for BCL2 translocations as biomarkers for their identification.

Figure 1.

Specific activity of NAMPTis in GCB-DLBCL. A, Plot of 211 repositionable drugs scored for DLBCL subtype specificity, expressed as the mathematical difference between the average AUC of GCB-DLBCL and the average AUC of ABC-DLBCL cell lines (n = 4 each; Y-axis). X-axis represents the drug activity in DLBCL, expressed as average AUC across the eight DLBCL cell lines. Representative drugs currently in use or in clinical trials for DLBCL are highlighted in green. The dotted line indicates a 30% ratio between activity and specificity. B, Dose–response curves to FK-866 in two ABC-DLBCL cell lines and two GCB-DLBCL cell lines, color coded as indicated. Error bars represent SDs (n = 6 points per measurement). Lines are fitted from a five-point logistic model. C, FK-866 IC₅₀ s for GCB- and ABC-DLBCL cell lines; GCB-S (dashed blue box) and GCB-R (dashed green box) subsets are indicated. Horizontal lines represent averages. P values were calculated using Student t test. *, P < 0.05.

Figure 2.

On-target activity of NAMPTis. A, Log IC₅₀ pairwise correlations between the indicated NAMPTis in 34 DLBCL cell lines (blue, GCB-DLBCL, n = 25; red, ABC-DLBCL, n = 9). Dotted lines represent linear regression models. R, Pearson correlation coefficient. B, Dose response of SUDHL4 to KPT-9724, alone (black) or with addition of 100 μmol/L NAMPT product β-NMN (green). C, AUC measurements of dose responses to each of the three NAMPTis, alone (black) or with addition of 100 μmol/L β -NMN (green), in three GCB-S DLBCL cell lines. D, KPT-9274 dose response in SUDHL4 cells transduced with EV (black), NAMPT WT (green) or NAMPT H191R cDNAs (red). E, AUC measurements of dose responses for three GCB-S DLBCL lines transduced with EV (black), NAMPT WT (green) or NAMPT H191R (red) cDNA and treated with each of the three NAMPTis. In the figure, error bars represent SDs; n = 6 (B and D) and 4 (C and E). P values were calculated using Student t test. ***, P < 0.001.

Figure 3.

NAMPT:PARP1 expression ratio determines NAMPTi sensitivity. A–C, NAMPT transcript levels (A), PARP1 transcript levels (B), and NP ratios (C) for the indicated subsets (GCB-S, blue, n = 13; GCB-R, green, n = 7; ABC, red, n = 9). Horizontal lines represent averages. D, Western blot analysis of V5-tagged NAMPT and vinculin expression in the indicated pools of sorted cells. H (high) and L (low) refer to Venus-sorted fractions. E, KPT-9274 dose–response curves in SUDHL4 cells transduced with the indicated UBC-driven lentiviral vectors and sorted in high or low fractions according to their average Venus fluorescence: EV low (EV-L, gray); EV high (EV-H, black); WT low (WT-L, light green); WT high (WT-H, dark green); H191R low (HR-L, orange); and H191R high (HR-H, red). F, KPT-9274 AUCs for the SUDHL4 pools indicated in D. G, Western blot analysis showing PARP1 suppression by the indicated sgRNAs. β-Tubulin, loading control. H and I, KPT-9274 dose responses of isogenic KARPAS (K)-422 (H) and WSU-DLCL2 (I) transduced with vectors expressing CAS9 and the indicated sgRNA (sgCtrl, black; sgPARP#1, green; sgPARP#2, orange). J and K, KPT-9724 dose response of KARPAS-422 (J) and WSU-DLCL2 (K) in the presence of increasing concentrations of the PARP inhibitor rucaparib. Rucaparib-only, green. Dotted line indicates 50% (KARPAS-422) or 75% (WSU-DLCL2) viability. L, KPT-9274 IC₅₀ s for KARPAS-422 cotreated with rucaparib at the indicated concentrations (μmol/L, RUCA). M, KPT-9274 IC₇₅ s for WSU-DLCL2 cotreated with rucaparib at the indicated concentrations (μmol/L, RUCA). Error bars represent SD (E, F, H–K) or SE (L and M); n = 6 (E, F, H, and I) and n = 4 (J, and K) measurements per points. P values were calculated using Student t test. *, P < 0.05; ***, P < 0.001; ABC vs. GCB comparisons for NAMPT and PARP1 transcript levels (A–C) are corrected for multiple hypothesis testing (see Supplementary Table S3). EV, empty vector; Ctrl, control; HR, NAMPT H191R; K, KARPAS; ns, not significant; RUCA, rucaparib; WT, NAMPT wild type.

Figure 4.

BCL2 translocation is a biomarker for NAMPTi sensitivity in GCB-DLBCL. A, Mean NP ratio in BCL2-translocated (n = 39, excluding double/triple hit) vs. not translocated (n = 94) GCB-DLBCL. B, Venn diagram of BCL2-translocated GCB-DLBCL and GCB-DLBCL cases in the bottom quartile of NP ratio distribution. C, Western blot analysis of the indicated proteins in WSU-DLCL2 and KARPAS-422 transduced with the indicated sgRNAs. D, Normalized GFP fraction (day 3 vs. day 0) of WSU-DLCL2 (right) and KARPAS-422 (left) cells carrying the indicated sgRNAs and left untreated (black bar) or treated with 500 nmol/L KPT-9274 for 3 days (red bar). E, Isobolograms for ABT-199 and KPT-9274 cotreatment in KARPAS-422 and WSU-DLCL2 (day 3). F, Heatmap of NAMPT and PARP1 gene expression levels and NP ratio in five GCB-DLBCL PDXs; BCL2 translocation status is indicated. G, Tumor volume measurement in PDX#20954 and PDX#75549 mice treated with KPT-9274 (black) or vehicle only (red) over the indicated times (n = 4/group). In the figure, error bars represent SDs. P values were calculated using Student t test with Welch correction (A), Fisher exact test (B), Student t test (D, E, and G). *, P < 0.05; **, P < 0.01; ***, P < 0.001. Ctrl, control; K, KARPAS; Neg, negative; ns, not significant; Pos, positive.