ALBERT

Description: Introduction. Targeting BCR signaling with the BTK inhibitor ibrutinib is clinically effective against most B-cell lymphomas, including the activated B-cell (ABC) subtype of diffuse large B-cell lymphoma (DLBCL), but not the germinal center B-cell (GCB) subtype. Active BCR signaling in GCB-DLBCL was suggested by studies with a Syk inhibitor and our previous studies using BCR knockout (KO). We addressed these questions: why is the BCR active in DLBCL, and how does it signal in GCB-DLBCL? Methods. We used CRISPR/Cas9 technology to modify selected genes by KO or homologous recombination-mediated knock-in (KI). For some genes KI was used to express a fluorescent protein (FP; e.g., GFP) instead of the targeted gene (KI/KO), or to modify the targeted gene together with KI of an FP, for detection of modified cells. Results. In GCB lines (OCI-Ly7 and OCI-Ly19) and ABC lines (U2932 and HBL-1), we simultaneously replaced the hypervariable region (HVR) exons of both immunoglobulin heavy (IgH) and light chains (IgL) with HVR sequences from normal B cells recognizing tetanus toxoid (TT). GFP and CFP respectively marked KI of IgH and IgL HVRs, and KI of the endogenous HVR sequences in each line served as controls. In CFP+/GFP+ cells, the TT specific BCR (TT-BCR) was expressed at similar or higher levels than the endogenous BCR (endo-BCR) and was functional, as shown by calcium flux in response to TT. The TT-BCR maintained growth of GCB lines (Fig. 1), indicating that they use "tonic", antigen-independent BCR signaling. Other features of tonic signaling were confirmed in more GCB lines: 1) the toxicity of BCR KO, which eliminates AKT S473 phosphorylation, was rescued by PTEN KO or expression of constitutively active AKT (mAKT), showing that BCR signaling serves principally to activate PI3K/AKT; and 2) KO of SYK or CD19, or truncation or ITAM mutation of the cytoplasmic tail of CD79A, none of which affect surface BCR levels, were as toxic as BCR KO but were non-toxic in BCR/PTEN double-KO cells. In contrast, the TT-BCR was as growth-slowing as BCR KO to the ABC line U2932 (Fig. 1), and substantially toxic to HBL-1, indicating that BCR signaling is self antigen-dependent in ABC-DLBCL. Reversion of somatic hypermutations in the U2932 HVRs was also as growth-slowing as BCR KO (Fig. 1), suggesting that self-antigen reactivity developed during BCR affinity maturation. Tonic signaling by the TT-BCR provided a detectable benefit (as compared to BCR KO) in PTEN-expressing HBL-1, whereas there was no difference between TT-HVR BCR and BCR KO in PTEN-deficient U2932. The surface TT-BCR level was higher than the endo-BCR level in ABC lines, and dropped with TT stimulation, suggesting that endo-BCRs in ABC lines undergo constant antigen stimulation with BCR internalization. The presumed self-antigen in ABC lines seems to be cell line-specific, since HVRs from ABC lines TMD8 and HBL-1 did not rescue growth of U2932. BCR KO in ABC lines was also not rescued by PTEN KO or mAKT. In cells whose BCRs were labeled by KI to fuse GFP to CD79A, super-resolution microscopy showed macro-clustering of BCR complexes at the surface of ABC line HBL-1, not seen in GCB lines (Fig. 2). Several findings suggested the clinical potential of targeting tonic BCR signaling in DLBCL: 1) clinical trial-stage inhibitors of SYK (P505-15) and PI3K (idelalisib) were toxic to GCB lines (less so with PTEN KO); 2) GCB lines (6/8) were sensitized by BCR KO to an in vitro CHOP-like regimen; 3) P505-15 or idelalisib sensitized GCB lines (3/3) to CHOP in vitro; and 4) evidence of tonic signaling in ABC line HBL-1 after removing antigen-driven signaling by HVR replacement. Conclusion. The BCR provides antigen-independent tonic signals to activate PI3K/AKT in GCB-DLBCL and antigen-dependent signaling in ABC-DLBCL. Targeting of B-cell specific tonic signling alone or in combination could be clinically effective in both types of DLBCL. Figure 1. Effect of BCR KO or HVR replacement in OCI-LY19 (A) and U2932 (B) cell lines. Endogenous IgH and IgL HVRs were replaced with HVR pairs (TT3 and/or TT6) recognizing tetanus toxoid, reverted to undo the effect of SHM, or restored with original HVRs. Figure 1. Effect of BCR KO or HVR replacement in OCI-LY19 (A) and U2932 (B) cell lines. Endogenous IgH and IgL HVRs were replaced with HVR pairs (TT3 and/or TT6) recognizing tetanus toxoid, reverted to undo the effect of SHM, or restored with original HVRs. Figure 2. Representative super-resolution images of BCR localization in live DLBCL cells. BCR labeled by CD79A-GFP fusion, surface membrane by CellMask staining. (bars = 5 µ m) Figure 2. Representative super-resolution images of BCR localization in live DLBCL cells. BCR labeled by CD79A-GFP fusion, surface membrane by CellMask staining. (bars = 5 µ m) Disclosures Westin: Spectrum: Research Funding.

Description: Introduction. Targeting antigen-driven B-cell receptor (BCR) signaling with the BTK inhibitor ibrutinib is clinically effective against most B-cell lymphomas, including activated B-cell diffuse large B-cell lymphoma (ABC-DLBCL), but not germinal center B-cell (GCB) DLBCL. We have formally confirmed that GCB-DLBCL cell lines utilize tonic BCR signaling, by showing: 1) sensitivity (variable) to knockout (KO) of the BCR, SYK, and CD19; 2) dependence on CD79A ITAM phosphorylation; and 3) independence from BCR antigen specificity. However, uncertainty remains about molecular events in upstream parts of tonic BCR signaling, why dependence of GCB-DLBCL cells on tonic BCR signaling is variable, and their clinical relevance. Methods. We used CRISPR/Cas9 methods to modify selected genes by KO and/or knock-in (KI) of the cDNA of a fluorescent protein (FP; e.g., GFP), with the FP serving as a marker of cells with gene KO or modification, or as a gene-fused tag for localization or quantitation. Cells expressing a membrane-targeted Forster resonance energy transfer (FRET) based AKT activity reporter (Lyn-AktAR2) were used to measure AKT activity directly by flow cytometry (FCM). Results. The effect of KI of CD79A Y188F mutation alone was similar to complete BCR KO, implying that CD79A Y188 phosphorylation is essential for tonic BCR signal transduction. Western blot analysis of GCB-DLBCL cell lines after BCR KO showed variable decreases of AKT S473 phosphorylation (frequently used as surrogate measure of AKT activity), but these did not correlate well with the variable decreases in proliferation of GCB-DLBCL cell lines caused by BCR KO. Measuring AKT activity directly (Fig. 1), or by another indirect approach (surface expression of CXCR4, a target gene of FOXO1 inhibited by AKT activity), showed high correlation between decreases in AKT activity and proliferation after BCR KO. In contrast to the variable effect of BCR KO on growth, pan-AKT KO was uniformly growth-slowing in GCB-DLBCL lines (Fig. 2). Interestingly, baseline surface density of BCR units in GCB lines, quantified by FCM using CD79A-GFP KI cells or anti-CD79B staining, correlated highly with reduction in growth or AKT activity caused by BCR KO (Fig. 3). These findings lead us to conclude that the BCR contributes to AKT activation in GCB-DLBCL cell lines, to a variable degree determined by BCR surface density. We also conclude that BCR surface density is determined by cell line-specific factors, as well as immunoglobulin heavy (IgH) and light (IgL) hypervariable region (HVR) sequences, based on measurements of BCR surface levels after exchanging endogenous HVR sequences in OCI-Ly19 and OCI-Ly7 cell lines for HVRs derived from other GCB and ABC-DLBCL cell lines. Reduction of AKT activity after BCR KO (measured by FRET reporter) and baseline BCR surface density in GCB-DLBCL cell lines also correlated well with the sensitivity of GCB-DLBCL lines to the clinically-tested SYK inhibitor (P505-15, PRT062607) or FDA-approved PI3K p110d isoform specific inhibitor (idelalisib). Interestingly, isogenic GCB-DLBCL cell lines with KO of PTEN, a negative regulator of AKT activation, were substantially more resistant to both inhibitors. A crucial role of PTEN deletion in overcoming dependence on tonic BCR signaling in GCB-DLBCL is supported by evidence from two naturally PTEN-deficient cell lines: SUDHL10, which adjusts to BCR KO and resumes normal growth, and HT, which lacks BCR expression, due to a frameshifting deletion in its IgH HVR. Re-expression of the BCR in HT, by KI to correct the IgH sequence, does not affect HT cell line growth. Conclusion. Our findings suggest a biomarker-guided therapeutic strategy in GCB-DLBCL: targeting tonic BCR signaling in BCR-high patients, by inhibiting CD79A phosphorylation, SYK, or PI3K, and downstream targeting of AKT in BCR-low and/or PTEN-deficient patients. Figure 1. Correlation of relative proliferation after BCR KO with decrease of AKT activity (as measured by FRET efficiency of AKT activity reporter) in GCB-DLBCL cell lines. Figure 1. Correlation of relative proliferation after BCR KO with decrease of AKT activity (as measured by FRET efficiency of AKT activity reporter) in GCB-DLBCL cell lines. Figure 2. Effect of BCR KO or pan-AKT KO in GCB-DLBCL cell lines. Figure 2. Effect of BCR KO or pan-AKT KO in GCB-DLBCL cell lines. Figure 3. Correlation of relative proliferation after BCR KO with baseline BCR surface density (as measured by flow cytometry of cells with CD79A-GFP fusion) in GCB-DLBCL cell lines. Figure 3. Correlation of relative proliferation after BCR KO with baseline BCR surface density (as measured by flow cytometry of cells with CD79A-GFP fusion) in GCB-DLBCL cell lines. Disclosures Burger: Pharmacyclics: Research Funding. Westin:Chugai: Membership on an entity's Board of Directors or advisory committees; Spectrum: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees; ProNAi: Membership on an entity's Board of Directors or advisory committees.

Description: Introduction. An essential role for the B-cell receptor (BCR) has been shown in multiple types of B-cell lymphoma by studies of cell lines and clinical responses to inhibitors of SYK or BTK. Diffuse large B-cell lymphoma (DLBCL) lines of the germinal center B-cell (GCB) type express a BCR, which can signal after crosslinking, but are unaffected by BCR pathway targeting toxic to lines of the activated B-cell (ABC) DLBCL subtype: knockdown of BCR signaling mediators (BTK, CD79A, and CD79B) by shRNA, and small-molecule inhibition of BTK by ibrutinib. GCB-DLBCL lines (and primary samples) also lack constitutive NF-kB activity and mutations in ITAM domains of CD79A or CD79B, BCR-related features of ABC-DLBCL. Most GCB-DLBCL patients resist BTK inhibition by ibrutinib, further suggesting that BCR signaling is not a feature of GCB-DLBCL. Methods. In 8 GCB-DLBCL lines (OCI-Ly7, OCI-Ly19, SUDHL-4, SUDHL-6, SUDHL-10, DB, BJAB, and HT) and one ABC-DLBCL line (HBL-1), we used electroporation to deliver a plasmid expressing Cas9 protein and a guide RNA (gRNA) targeting one of these: constant exons of IGHM, IGHG, or Igκ; the cell line-specific IgH hypervariable region (HVR); or CXCR4. Knock-in (KI) of mouse CD8a (mCD8a), after the HVR V segment leader sequence and followed by a polyA signal, was used as a positive marker of BCR knockout (KO) in HBL-1 and OCI-Ly19 cell lines. Surface BCR, CXCR4, and mCD8a were detected by flow cytometry (FACS). BCR KO cells were viably sorted 4-6 days after electroporation, cultured 1-3 days more, and studied by whole-genome gene expression profiling (GEP) on Illumina HT12v4 arrays and Western blotting. Results. Only 2 days after electroporation, FACS showed cells with correlated loss of surface BCR proteins (IgH, Igκ or Igl, and CD79B), which eventually declined to undetectable levels. Forward and side scatter showed that BCR KO cells were smaller. The proportion of BCR KO (or mCD8a KI/KO) cells declined over time, steadily after complete BCR elimination (Fig. 1A). BCR KO cells in GCB-DLBCL lines grew more slowly than BCR-replete cells but variably, from almost no difference in BJAB to growth cessation in SUDHL-4, SUDHL-10 and HBL-1 (Fig. 1B). CXCR4 KO cells were a stable proportion (Fig. 1A) with a normal growth rate (Fig. 1B), indicating that growth reduction by BCR KO is specific. Continued expression of mCD8a indicated viability and sustained IgH transcription in BCR KO cells. Cell cycle analysis showed lower proportions of S and G2/M phases in BCR KO cells, proportional to growth retardation, and sub-G1 cells in OCI-Ly7 (Fig. 2), SUDHL-4 and SUDHL-10. Apoptosis in OCI-Ly7 BCR KO cells was confirmed with a caspase-3 fluorogenic substrate. Igκ KO similarly caused complete BCR loss and growth retardation, in OCI-Ly7 cells even more than with IgH KO. In the HT cell line, which lacked BCR expression due to a single-nucleotide deletion in its IgH HVR, KI repaired the HVR and caused expression of surface BCR (IgM with Igκ and CD79B) but no change in growth rate, suggesting BCR-proximal activators of BCR signaling pathways. Targeted BCR KO is not currently a therapeutic option, but BCR KO cells were relatively more sensitive to an in vitro regimen modeling the non-prednisone drugs of CHOP. No change in drug sensitivity was observed with BCR KO in BJAB, or in CXCR4 KO cells. GEP showed that BCR KO downregulated several genes characteristically expressed by GCB-DLBCL, and genes associated with negative regulation of BCR signaling. Pathway analysis with Gene Set Enrichment Analysis (GSEA) showed that BCR KO reduced expression of proliferation-related signatures, and produced changes associated with B-cell differentiation stages lacking a mature BCR, either early (pre-B cells) or late (plasma cells). GSEA implicated loss of MAPK/ERK and PI3K/AKT signaling pathways as mediators of BCR KO-induced changes, confirmed by Western blotting showing loss of phosphorylation of SYK, AKT and ERK after BCR KO. Conclusions. Complete BCR KO by Cas9/gRNA showed that GCB-DLBCL lines require the BCR for optimal viability, cell growth, and chemotherapy resistance. BCR KO-induced changes are mediated by MAPK/ERK and PI3K/AKT signaling pathways. Table A. B. Figure 1. Figure 1A. BCR KO cells (distinguished from BCR-replete cells by FACS), but not CXCR4 KO cells, show relative decline (A) and slower absolute growth (B) in mixed cultures. Figure 1A. BCR KO cells (distinguished from BCR-replete cells by FACS), but not CXCR4 KO cells, show relative decline (A) and slower absolute growth (B) in mixed cultures. Figure 1B Figure 1B. Figure 2 Cell cycle changes with BCR KO in OCI-Ly7. Figure 2. Cell cycle changes with BCR KO in OCI-Ly7. Disclosures No relevant conflicts of interest to declare.

Description: Key Points The GCB subtype of DLBCL relies exclusively on tonic BCR signaling via CD79A Y188. PTEN protein expression and BCR surface density determine the contribution of tonic BCR signaling to AKT activity in GCB-DLBCL.