Inhibition of PRMT5 sensitizes B-cell non-Hodgkin lymphoma cells to both intrinsic and extrinsic apoptotic cell death∗ Authors: Fen Zhu, Jing Wang, Christine E. Ryan, Liam Hackett, Jeremy Zhang, Stephanie Chamberlain, Johnson Ung, Stephen J. F. Chong, Matthew S. Davids Published: 2026 ABSTRACT Key Points•JNJ-64619178 increases mitochondrial apoptotic priming and synergizes with venetoclax to induce apoptosis in B-NHL cells.•JNJ-64619178 upregulates cell-surface TRAIL receptors (DR4/DR5) and enhances TRAIL-mediated apoptosis in B-NHL cells. Introduction Protein arginine methylation, catalyzed by protein arginine methyltransferases (PRMTs) using S-adenosyl methionine as methyl group donor, is a common posttranslational protein modification that alters the protein interactome by increasing residue bulk and accessibility.1 PRMTs catalyze monomethylation followed by asymmetric or symmetric dimethylation of arginine, designated as type I or type II arginine methylation, respectively, which may lead to different biological functions.2 PRMT5, a type II methyltransferase, targets a broad set of substrates such as histones, splicing factors, transcription factors, and other signaling molecules, thereby influencing many essential cellular processes.3 Histones (H2A, H3, and H4) and spliceosome proteins (SmD1, SmD3, and SmB) are established PRMT5 substrates, and PRMT5 inhibition disrupts gene transcription and splicing.4,5 Methylation of transcription factors and signaling molecules appears cell-type dependent, further expanding the functional scope of PRMT5.3 PRMT5 expression is increased in diffuse large B-cell lymphoma (DLBCL), the most common subtype of aggressive B-cell non-Hodgkin lymphoma (B-NHL), compared with naïve B cells because of activation of B-cell receptor signaling.6 PRMT5 inhibition delays DLBCL cell proliferation both in vitro and in vivo.6 Mantle cell lymphoma (MCL), an incurable B-NHL derived from a subset of naïve pregerminal center B cells, also depends on PRMT5 for survival, and previous studies have explored the activity of novel small-molecule inhibitors targeting PRMT5 in MCL.7,8 In relapsed/refractory MCL, PRMT5 expression is upregulated and positively associated with poor prognosis, and inhibition of PRMT5 delays the growth of chimeric antigen receptor (CAR) T-cell relapsed MCL cells.9 As such, small-molecule inhibitors targeting PRMT5 warrant further exploration in DLBCL and MCL. JNJ-64619178 is a small-molecule inhibitor of PRMT5 that occupies both the S-adenosyl methionine and substrate-binding pockets of PRMT5, thereby reducing substrate methylation. It shows broad antiproliferative activity across solid and hematological cancer cell lines.10 In a phase 1 trial in lower-risk myelodysplastic syndromes, JNJ-64619178 demonstrated mainly manageable hematological toxicity but no clinical responses.11 A separate phase 1 study in advanced solid tumors and NHL reported dose-dependent, manageable toxicity, and preliminary antitumor activity of JNJ-64619178,12 warranting further exploration of rational combination strategies. Dysregulation of both the mitochondrial (intrinsic) and extrinsic apoptotic pathways contributes to cancer pathogenesis. The B-cell lymphoma (BCL) 2 inhibitor venetoclax has demonstrated clinical efficacy in certain B-NHLs,13 particularly chronic lymphocytic leukemia, for which venetoclax was first approved by the US Food and Drug Administration. However, venetoclax monotherapy efficacy beyond chronic lymphocytic leukemia has been modest, prompting the need to develop combination approaches. To identify rational BH3 mimetic partners for JNJ-64619178 in B-NHL, we used dynamic BH3 profiling (DBP), a functional assay that measures the net proapoptotic effect of drugs of interest in cancer cells, including the proximity of the cancer cells to the threshold of apoptosis (mitochondrial apoptotic priming) and the dependence of cancer cells on individual antiapoptotic BCL-2 family proteins.14,15 Extrinsic apoptosis is triggered by activation of death receptors such as Fas, tumor necrosis factor receptor-1, and TRAIL receptors DR4 and DR5. TRAIL binding to its receptor DR4/DR5 leads to receptor trimerization and subsequent formation of the death-inducing signaling complex.16 TRAIL-mimetic agents that activate DR4/DR5 can selectively induce cancer cell apoptosis while sparing normal cells.17,18 Although clinical trials of TRAIL derivatives and DR4/DR5 agonists demonstrated favorable safety, efficacy in solid tumors was limited.19, 20, 21, 22, 23 These findings indicate that agents targeting extrinsic apoptosis may be more effective when used in biologically informed combinations. In this study, we used a 2-pronged approach to identify potential synthetic lethality induced by JNJ-64619178, which included DBP to assess how JNJ-64619178 alters mitochondrial apoptotic priming and antiapoptotic dependence in DLBCL and MCL cells, thereby guiding selection of combination partners, and RNA sequencing (RNA-seq) with pathway enrichment analysis to examine how JNJ-64619178 reshapes signaling pathways to resensitize cells to additional therapies. Our findings highlight the role of PRMT5 in regulating both intrinsic and extrinsic apoptosis and identify potential therapeutic combinations between PRMT5 inhibitors and apoptosis-inducing agents in B-NHL. Methods Cell lines DLBCL cell lines (TMD8, OCI-Ly3, OCI-Ly1, and SUDHL4), venetoclax-resistant OCI-Ly1 cell line (OCI-Ly1R),24 Burkitt Lymphoma (BL) cell lines (Raji, BL-30, and BL-70), and MCL cell lines (Mino, Jeko-1) were cultured in RPMI 1640 media supplemented with 10% fetal bovine serum, 100 U/mL penicillin-streptomycin, 2mM L-glutamine. OCI-Ly1R was maintained with 0.5μM venetoclax. HEK293T cells were cultured in Dulbecco (or Dulbecco’s) modified Eagle medium supplemented with 10% fetal bovine serum and 100 U/mL penicillin-streptomycin. All cell lines were authenticated using short tandem repeat profiling and tested negative for mycoplasma. The double-hit lymphoma patient-derived xenograft (PDX) cell line, DW19, was cultured in X-VIVO 15 medium supplemented with 20 ng/mL fungizone and isolated from PDX mice available from the Public Repository of Xenografts. All cultures were maintained at 37°C in 5% CO2. Xenografts All murine studies were performed according to protocols approved by the Dana-Farber Cancer Institute Institutional Animal Care and Use Committee. NOD/SCID/Il2Rγ−/− (NSG) mice were purchased from the Jackson Laboratory. Luciferase-labeled OCI-Ly1 cells were injected through the tail veins into male mice aged 7 weeks. Bioluminescent imaging (BLI) confirmed engraftment 5 days later, and mice were randomized to treatment groups with comparable tumor burden. For studies on the combination of JNJ-64619178 and venetoclax, mice were assigned to 4 cohorts: (1) vehicle; (2) JNJ-64619178; (3) venetoclax; and (4) the combination cohorts. Similarly, for studies on the combination of JNJ-64619178 and conatumumab, mice received (1) vehicle; (2) JNJ-64619178; (3) conatumumab; or (4) the combination treatment. After treatments, mice were monitored until they reached end point criteria (eg, weight loss, paralysis). Tumor burden was monitored weekly by BLI. BH3 profiling BH3 profiling was performed by flow cytometry, as described previously.15 Briefly, cells were exposed to digitonin and BH3-only peptides or mimetics for 60 minutes in MEB2 buffer. After formaldehyde fixation and neutralization, fluorescent antibodies Alexa Fluor 488 anticytochrome c and Hoechst 33342 were added. Cells were incubated with antibodies for at least 24 hours at 4°C before being analyzed by flow cytometry. BH3 profiling data were presented as a percentage of cytochrome c loss. Statistical analysis Statistical analyses were conducted using Prism 10 (GraphPad Software). A P value of <.05 was considered significant. The specific comparison methods used in each experiment are detailed in the figure legends. Drug synergy was calculated using SynergyFinder Plus.25 Densitometry of western blot protein bands was quantified using ImageJ.26 Further information can be found in the supplemental Materials. Results PRMT5 inhibition by JNJ-64619178 controls proliferation and transcription in B-NHL To investigate how JNJ-64619178 affects B-NHL proliferation, we treated DLBCL (n = 4), MCL (n = 2), and BL (n = 2) cell lines with 0.2μM to 5μM JNJ-64619178 for 5 days and measured cell proliferation by CellTiter-Glo. OCI-Ly1R was the most sensitive, followed by its parental OCI-Ly1, whereas other lines showed variable, moderate responses (Figure 1A).Figure 1.JNJ-64619178 suppresses the proliferation of B-NHL cell lines and globally affects gene expression and splicing. (A) A panel of B-NHL cell lines was treated with JNJ-64619178 at indicated dosages for 5 days. Cell proliferation was measured by CellTiter-Glo assay. Data are presented as mean ± standard deviation (SD), n = 3. (B-C) OCI-Ly1 and SUDHL4 cell lines were treated with DMSO or 1μM JNJ-64619178 for 5 days. (B) Cell viability was measured by propidium iodide (PI) and annexin V staining, data are presented as mean ± SD, n = 3. (C) Cell cycle was determined by 5-bromo-2′-deoxyuridine (BrdU) and 7-aminoactinomycin D (7-AAD) cell cycle analysis kit. P value for each cell cycle phase was calculated using Student t test by comparing DMSO treatment with JNJ-64619178 treatment, n = 3. (D-F) RNA-seq analysis of SUDHL4 treated with DMSO or 1μM JNJ-64619178 for 3 days. (D) Top 20 significantly enriched pathways in JNJ-64619178 treated cells compared with DMSO treatment. (E) Distribution of differentially spliced events in JNJ-64619178 treated SUDHL4 compared with DMSO treatment. (F) Sashimi plot showing exon 6 skipping in MDM4 gene in JNJ-64619178 treated SUDHL4. A3SS, alternative 3′ splice sites; A5SS, alternative 5′ splice sites; EBV, Epstein–Barr Virus; JNJ-9178, JNJ-64619178; MXE, mutually exclusive exons; RI, retained introns; SE, skipped exons. JNJ-64619178 suppresses the proliferation of B-NHL cell lines and globally affects gene expression and splicing. (A) A panel of B-NHL cell lines was treated with JNJ-64619178 at indicated dosages for 5 days. Cell proliferation was measured by CellTiter-Glo assay. Data are presented as mean ± standard deviation (SD), n = 3. (B-C) OCI-Ly1 and SUDHL4 cell lines were treated with DMSO or 1μM JNJ-64619178 for 5 days. (B) Cell viability was measured by propidium iodide (PI) and annexin V staining, data are presented as mean ± SD, n = 3. (C) Cell cycle was determined by 5-bromo-2′-deoxyuridine (BrdU) and 7-aminoactinomycin D (7-AAD) cell cycle analysis kit. P value for each cell cycle phase was calculated using Student t test by comparing DMSO treatment with JNJ-64619178 treatment, n = 3. (D-F) RNA-seq analysis of SUDHL4 treated with DMSO or 1μM JNJ-64619178 for 3 days. (D) Top 20 significantly enriched pathways in JNJ-64619178 treated cells compared with DMSO treatment. (E) Distribution of differentially spliced events in JNJ-64619178 treated SUDHL4 compared with DMSO treatment. (F) Sashimi plot showing exon 6 skipping in MDM4 gene in JNJ-64619178 treated SUDHL4. A3SS, alternative 3′ splice sites; A5SS, alternative 5′ splice sites; EBV, Epstein–Barr Virus; JNJ-9178, JNJ-64619178; MXE, mutually exclusive exons; RI, retained introns; SE, skipped exons. Next, we used OCI-Ly1 and SUDHL4 for mechanistic studies. Despite differing sensitivities in CellTiter-Glo assays, both cell lines exhibited comparable levels of apoptosis after JNJ-64619178 treatment, suggesting metabolic rather than cytotoxic differences (Figure 1B). Cell cycle analysis revealed reduced S-phase and G2/M-phase accumulation in both cell lines in response to JNJ-64619178 treatment (Figure 1C; supplemental Figure 1A). To define PRMT5-dependent transcriptional programs, we performed RNA-seq on OCI-Ly1 and SUDHL4 cells treated with dimethyl sulfoxide (DMSO) or 1μM JNJ-64619178 for 3 days. Pathway analysis showed enrichment of several biological pathways, with spliceosome, lysosome, endocytosis, and apoptosis pathways being enriched in both cell lines (Figure 1D; supplemental Figure 1B). Although the role of PRMT5 in gene splicing is established, the functions of most of its splicing target genes remain uncharacterized.5,8,27,28 Transcriptomic analysis identified exon skipping as the predominant splicing alteration in response to JNJ-64619178, including exon 6 skipping in the MDM4 gene in both cell lines (Figure 1E-F; supplemental Figure 1C), a known marker of PRMT5 inhibition.6,8,28 PRMT5 inhibition has also been linked to aberrant splicing of several BCL-2 family members, although the functional significance of these events remains unclear.8 In our analysis, JNJ-64619178 induced exon 2 skipping in BAK1 (supplemental Figure 1D-E), confirmed by polymerase chain reaction (supplemental Figure 1F). Exon 2 is located in the 5′ untranslated region, and skipping of this exon does not change the protein sequence. However, because BAK is an effector proapoptotic protein, we asked whether exon 2 might modulate BAK function by influencing mRNA stability or translation. To test this, we generated myc-tagged BAK1 constructs expressing distinct transcript variants and overexpressed them in HEK293T cells. Neither BAK protein abundance nor RNA stability was affected by exon 2 status (supplemental Figure 1G-H), suggesting minimal functional relevance. Collectively, these data show that JNJ-64619178 suppresses proliferation, induces apoptosis and cell cycle arrest, and broadly disrupts transcriptional and splicing programs in B-NHL cell lines, supporting further exploration of combination strategies to enhance its therapeutic efficacy. JNJ-64619178 increases mitochondrial apoptotic priming and synergizes with BH3 mimetics in inducing apoptosis We next investigated combination strategies to enhance the cytotoxicity of JNJ-64619178. We observed that mRNA levels of BCL-2 family members were not significantly altered after JNJ-64619178 treatment (supplemental Figure 2A), and protein expression showed minimal or inconsistent changes among cell lines and across replicates (supplemental Figure 2B-C). However, previous studies have shown that shifts in interactions among proapoptotic and antiapoptotic proteins, as well as their posttranslational modifications, can strongly influence mitochondrial apoptotic sensitivity independent of expression levels.29,30 To investigate the effects of JNJ-64619178 on mitochondrial apoptosis and to identify proapoptotic drug combinations, we performed DBP using synthetic BH3 peptides and mimetics with distinct binding preferences for antiapoptotic BCL-2 family proteins (Figure 2A). JNJ-64619178 increased mitochondrial apoptotic priming and BCL-2 dependence in DLBCL and MCL cell lines (Figure 2B). Some cell lines also showed elevated reliance on BCL-xL, MCL-1, or B-cell lymphoma 2–related protein A1 (BFL-1). Notably, these changes were observed as early as 3 days, before measurable effects on apoptosis, indicating that modulation of mitochondrial apoptotic tendency is an early cellular response to JNJ-64619178 in DLBCL and MCL cell lines.Figure 2.JNJ-64619178 increases mitochondrial apoptotic priming and synergizes with BH3 mimetics in inducing apoptosis. (A) (Left) DBP workflow in which cells treated with vehicle control or drug of interest are incubated with BH3 peptides/mimetic to induce cytochrome c release as measured by flow cytometry. (Right) Interaction map between BH3 peptides/mimetic and antiapoptotic proteins. Blue indicates a high-affinity interaction. (B) DBP in the indicated cell lines upon 3 days of treatment with DMSO or 1μM JNJ-64619178. (C-D) Cell lines were treated as indicated with JNJ-64619178 for 5 days, and venetoclax, S63845, or A1331852 were added on the last 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. (E) Summary of mean Loewe synergy scores from a panel of DLBCL and MCL cell lines treated with the combination of JNJ-64619178 and BH3 mimetics. (F) Summary of maximum cell killing by the indicated drug combination in DLBCL and MCL cell lines. FACS, Fluorescence-Activated Cell Sorting; Max, maximum; ven, venetoclax. JNJ-64619178 increases mitochondrial apoptotic priming and synergizes with BH3 mimetics in inducing apoptosis. (A) (Left) DBP workflow in which cells treated with vehicle control or drug of interest are incubated with BH3 peptides/mimetic to induce cytochrome c release as measured by flow cytometry. (Right) Interaction map between BH3 peptides/mimetic and antiapoptotic proteins. Blue indicates a high-affinity interaction. (B) DBP in the indicated cell lines upon 3 days of treatment with DMSO or 1μM JNJ-64619178. (C-D) Cell lines were treated as indicated with JNJ-64619178 for 5 days, and venetoclax, S63845, or A1331852 were added on the last 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. (E) Summary of mean Loewe synergy scores from a panel of DLBCL and MCL cell lines treated with the combination of JNJ-64619178 and BH3 mimetics. (F) Summary of maximum cell killing by the indicated drug combination in DLBCL and MCL cell lines. FACS, Fluorescence-Activated Cell Sorting; Max, maximum; ven, venetoclax. We previously showed that most DLBCL cells resist single-agent BH3 mimetics because of simultaneous dependence on multiple antiapoptotic proteins.29 DBP suggested that combining JNJ-64619178 with BH3 mimetics could enhance cytotoxicity. Therefore, we combined JNJ-64619178 with venetoclax (BCL-2i), S63845 (MCL-1i), or A1331852 (BCL-xLi) in several B-NHL cell lines. Cells were pretreated with JNJ-64619178 for 3 days to induce apoptotic priming, followed by 2 days of combination treatment. Viability was assessed by PI/annexin V assay (Figure 2C-D; supplemental Figure 3). Synergy analysis (mean Loewe score ≥10) showed that JNJ-64619178 synergized with venetoclax across all 7 DLBCL and MCL cell lines, whereas the combination with S63845 or A1331852 was synergistic in 4 of 7 cell lines (Figure 2E). Maximum cell killing was achieved with JNJ-64619178/venetoclax in 5 cell lines and with JNJ-64619178/S63845 in 2 cell lines (Figure 2F), indicating that venetoclax is the most broadly effective partner for JNJ-64619178. We next evaluated caspase activation, a hallmark of apoptosis, after JNJ-64619178 and venetoclax cotreatment. Western blot showed enhanced caspase-3 cleavage after cotreatment in TMD8 cells (supplemental Figure 4A). Caspase-9 activity, measured by caspase-glo, was also significantly increased after JNJ-64619178 and venetoclax treatment in OCI-Ly1, SUDHL4, and Mino cells (supplemental Figure 4B). The pan-caspase inhibitor Z-VAD-FMK markedly reduced combination-induced cytotoxicity (supplemental Figure 4C), confirming a caspase-dependent mechanism. Collectively, these findings show that JNJ-64619178 induces BCL-2 dependence and that cotreatment with venetoclax drives synergistic, caspase-dependent apoptosis. JNJ-64619178 does not shift antiapoptotic dependence to BCL-2 in less BCL-2–dependent cells The DLBCL and MCL cell lines used thus far display baseline BCL-2 dependence, sometimes with codependence on MCL-1, BCL-xL, or BFL-1. We next asked whether JNJ-64619178 could shift antiapoptotic dependence toward BCL-2 in cancer cells with lower baseline BCL-2–dependence, thereby sensitizing them to venetoclax. To test this, we evaluated 3 cell lines with negligible baseline BCL-2 dependence driven by distinct mechanisms. The first was Raji, which primarily depends on MCL-1 and BFL-1 at baseline (Figure 3A). JNJ-64619178 exposure increased overall apoptotic priming and strengthened dependence on MCL-1 and BFL-1, but not BCL-2 (Figure 3A). Despite detectable BCL-2 protein expression in Raji, functional BCL-2 dependence was absent (supplemental Figure 5). The second cell line, BL-70, lacks BCL-2 protein (supplemental Figure 5) and shows minimal BCL-2 dependence (Figure 3B). BL-70’s baseline MCL-1 dependence was already maximal, but JNJ-64619178 treatment nonetheless increased overall priming and dependence on BCL-xL and BFL-1 (Figure 3B). Third, OCI-Ly1R, a venetoclax-acquired resistant cell line, displays increased MCL-1 and decreased BCL-2 dependence compared with the parental OCI-Ly1 line because of altered expression and posttranslational regulation of BCL-2 family proteins.30 Treatment with JNJ-64619178 elevated overall priming as well as dependence on BCL-xL, BFL-1, and MCL-1, but again failed to induce BCL-2 dependence (Figure 3C). Collectively, these results show that all 3 cell lines are strongly MCL-1-dependent and have acquired venetoclax resistance through different mechanisms, and that JNJ-64619178 does not reprogram their antiapoptotic dependence toward BCL-2.Figure 3.JNJ-64619178 does not shift apoptotic dependence to BCL-2 in less BCL-2–dependent cells. (A-C) DBP in Raji, BL-70, and OCI-Ly1R cells after 3 days of treatment with DMSO or 1μM JNJ-64619178. (D-F) Raji, BL-70, and OCI-Ly1R cells were treated as indicated with JNJ-64619178 for 5 days; venetoclax and S63845 were added on the last 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. JNJ-64619178 does not shift apoptotic dependence to BCL-2 in less BCL-2–dependent cells. (A-C) DBP in Raji, BL-70, and OCI-Ly1R cells after 3 days of treatment with DMSO or 1μM JNJ-64619178. (D-F) Raji, BL-70, and OCI-Ly1R cells were treated as indicated with JNJ-64619178 for 5 days; venetoclax and S63845 were added on the last 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. DBP suggested that JNJ-64619178 might further sensitize these cell lines to MCL-1 rather than BCL-2 targeted therapy. To test this, we treated these cell lines with JNJ-646191678, venetoclax, and S63845 alone or in combination and assessed cell-killing synergy. Raji was resistant to JNJ-64619178 and venetoclax (Figure 3D), and the combination failed to activate caspases (supplemental Figure 4A-B). In contrast, S63845 induced apoptosis in Raji and synergized with JNJ-64619178 to further enhance apoptosis (Figure 3D). BL-70 was sensitive to JNJ-64619178-induced apoptosis but resistant to venetoclax. Cotreatment with venetoclax did not increase killing beyond that achieved by JNJ-64619178 alone (Figure 3E). Among the 3 cell lines, BL-70 exhibited the strongest MCL-1 dependence by BH3 profiling and showed the greatest sensitivity to S63845 monotherapy, with low concentrations sufficient to induce apoptosis. This sensitivity was further enhanced when combined with JNJ-64619178 (Figure 3E). OCI-Ly1R was more sensitive to JNJ-64619178 than its parental OCI-Ly1 counterpart (Figure 1A). However, even with priming by JNJ-64619178, venetoclax induced apoptosis only at concentrations far higher than those effective in BCL-2–dependent cell lines. OCI-Ly1R was sensitive to S63845 as a single agent, and its response was augmented by cotreatment with JNJ-64619178 (Figure 3F). Thus, JNJ-64619178 increased mitochondrial apoptotic priming in OCI-Ly1R without shifting antiapoptotic dependence from MCL-1 to BCL-2. Overall, these findings indicate that although JNJ-64619178 enhances mitochondrial apoptotic priming across diverse B-NHL cell lines, it does not alter their intrinsic antiapoptotic dependence. JNJ-64619178 and venetoclax combination is active in vivo in a DLBCL cell line–derived xenograft mouse model We established a xenograft mouse model using luciferase-expressing OCI-Ly1 cells to evaluate the in vivo activity of the combination of JNJ-64619178 and venetoclax. Upon confirming successful engraftment by BLI, mice were assigned into 4 groups with comparable tumor burden treated starting 1 day later with vehicle control, JNJ-64619178, venetoclax, or the combination treatment. Both drugs were well tolerated by mice at the administered doses. Without any treatment, tumors progressed rapidly, and all control mice developed paralysis within 3 weeks of engraftment. Monotherapy with either agent significantly delayed tumor growth and prolonged survival compared with controls; however, the combination provided superior benefit by more effectively suppressing tumor growth and prolonging survival (Figure 4A-C). Although single agents temporarily halted tumor growth, tumor burden progressed rapidly thereafter (Figure 4A-B), suggesting limited single-agent efficacy in vivo. In contrast, the combination of JNJ-64619178 and venetoclax suppressed tumor growth throughout the duration of therapy. Together, these findings demonstrate that the combination of JNJ-64619178 and venetoclax effectively suppresses tumor growth in vivo.Figure 4.JNJ-64619178 and venetoclax delay tumor growth and prolong survival in a DLBCL cell line-derived xenograft mouse model. NSG mice transplanted with luciferase-expressing OCI-Ly1 were assigned into 4 groups to receive either vehicle control (n = 5), venetoclax (n = 5), JNJ-64619178 (n = 5), or venetoclax + JNJ-64619178 (n = 5). (A) Representative BLI from each group. The first BLI was performed on day 0, which is 5 days after injection of tumor cells. Treatment started on day 1. (B) Quantification of BLI signal over time. Graph shows mean ± standard error of the mean. P value was calculated using Student t test by comparing the combination group with other groups. (C) Kaplan-Meier survival curve showing the percentage of mice without symptoms of disease (weight loss, paralysis). P value was calculated by log-rank test. Min, minimum. JNJ-64619178 and venetoclax delay tumor growth and prolong survival in a DLBCL cell line-derived xenograft mouse model. NSG mice transplanted with luciferase-expressing OCI-Ly1 were assigned into 4 groups to receive either vehicle control (n = 5), venetoclax (n = 5), JNJ-64619178 (n = 5), or venetoclax + JNJ-64619178 (n = 5). (A) Representative BLI from each group. The first BLI was performed on day 0, which is 5 days after injection of tumor cells. Treatment started on day 1. (B) Quantification of BLI signal over time. Graph shows mean ± standard error of the mean. P value was calculated using Student t test by comparing the combination group with other groups. (C) Kaplan-Meier survival curve showing the percentage of mice without symptoms of disease (weight loss, paralysis). P value was calculated by log-rank test. Min, minimum. JNJ-64619178 upregulates cell surface TRAIL receptors and enhances TRAIL-induced apoptosis Beyond intrinsic mitochondria-dependent apoptosis, we examined extrinsic apoptosis through death receptor signaling. RNA-seq revealed that DR5 is the most abundantly expressed apoptosis-inducing death receptor in DLBCL cell lines, followed by DR4. Both receptors serve as TRAIL receptors and were not significantly altered transcriptionally by JNJ-64619178 (supplemental Figure 6A). JNJ-64619178 treatment significantly increased cell surface DR4 and DR5 across DLBCL, MCL, and BL cell lines (Figure 5A-C; supplemental Figure 6B-E), an effect not observed with venetoclax (supplemental Figure 6F-G). Because DR5 upregulation can enhance sensitivity to TRAIL,31 we asked whether these changes enhance responsiveness to TRAIL. Although baseline responses varied, JNJ-64619178 synergized with TRAIL to induce apoptosis in every cell line tested, including the less BCL-2–dependent cell lines (Figure 5D-G; supplemental Figure 6H-M). These results demonstrate that JNJ-64619178 upregulates cell surface DR4/DR5 expression and broadly sensitizes B-NHL cell lines to TRAIL-induced apoptosis.Figure 5.JNJ-64619178 upregulates DR4 and DR5 cell surface expression and sensitizes B-NHL cell lines to TRAIL-induced apoptosis. (A-C) SUDHL4, Raji, and Mino cells were treated with DMSO or 1μM JNJ-64619178 for 5 days; DR4 and DR5 expression level on cell membrane of live cells (PI-negative) were analyzed via flow cytometry. (Top) Representative histograms. (Bottom) Quantification of median fluorescent intensity (MFI) normalized to DMSO treatment. P value was calculated by Student t test. Data are presented as mean ± SD, n = 3. (D-F) SUDHL4, Raji, and Mino cells were treated with DMSO or 1μM JNJ-64619178 for 5 days, and recombinant human TRAIL was added on the last 2 days at indicated dosages. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. (G) Summary of mean Loewe synergy scores from a panel of BL, DLBCL, and MCL cell lines treated with the combination of JNJ-64619178 and TRAIL as previously described. (H) SUDHL4 were treated with 1μM H3B-8800 for 2 days, DR4 and DR5 expression levels on cell membrane were analyzed via flow cytometry, normalized MFI was shown as mean ± SD, n = 3. P value was calculated using Student t test. (I) SUDHL4 were treated with the combination of 1μM H3B-8800 and recombinant human TRAIL for 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. (J) SUDHL4 were treated with 20μM E7070 for 2 days, DR4, and DR5 expression levels on cell membrane were analyzed via flow cytometry, MFI was shown as mean ± SD, n = 3. P value was calculated using Student t test. (K) SUDHL4 were treated with the combination of 20μM E7070 with recombinant human TRAIL for 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. JNJ-64619178 upregulates DR4 and DR5 cell surface expression and sensitizes B-NHL cell lines to TRAIL-induced apoptosis. (A-C) SUDHL4, Raji, and Mino cells were treated with DMSO or 1μM JNJ-64619178 for 5 days; DR4 and DR5 expression level on cell membrane of live cells (PI-negative) were analyzed via flow cytometry. (Top) Representative histograms. (Bottom) Quantification of median fluorescent intensity (MFI) normalized to DMSO treatment. P value was calculated by Student t test. Data are presented as mean ± SD, n = 3. (D-F) SUDHL4, Raji, and Mino cells were treated with DMSO or 1μM JNJ-64619178 for 5 days, and recombinant human TRAIL was added on the last 2 days at indicated dosages. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. (G) Summary of mean Loewe synergy scores from a panel of BL, DLBCL, and MCL cell lines treated with the combination of JNJ-64619178 and TRAIL as previously described. (H) SUDHL4 were treated with 1μM H3B-8800 for 2 days, DR4 and DR5 expression levels on cell membrane were analyzed via flow cytometry, normalized MFI was shown as mean ± SD, n = 3. P value was calculated using Student t test. (I) SUDHL4 were treated with the combination of 1μM H3B-8800 and recombinant human TRAIL for 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. (J) SUDHL4 were treated with 20μM E7070 for 2 days, DR4, and DR5 expression levels on cell membrane were analyzed via flow cytometry, MFI was shown as mean ± SD, n = 3. P value was calculated using Student t test. (K) SUDHL4 were treated with the combination of 20μM E7070 with recombinant human TRAIL for 2 days. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. Next, we investigated the mechanism underlying DR4/DR5 upregulation by JNJ-64619178. Golgi or endoplasmic reticulum (ER) stress has been reported to increase DR4 or DR5 expression and promote apoptosis.32,33 Because PRMT5 localizes to the Golgi and maintains its structure through GM130 methylation, its inhibition could disrupt Golgi integrity and induce stress.34 However, total DR4 and DR5 levels showed no consistent change after JNJ-64619178 treatment across cell lines (supplemental Figure 6N), arguing against Golgi- or ER-stress–mediated upregulation. Thus, increased cell surface DR4 and DR5 expression after JNJ-64619178 treatment is unlikely because of changes in total protein level. Given that JNJ-64619178 induced widespread splicing alterations, we next examined whether splicing perturbation could drive increased cell surface DR expression. To test this, we used 2 splicing modulator drugs, H3B-8800 and E7070, which target SF3B1 and RBM39, respectively, within the splicing complex. H3B-8800 has been shown to preferentially kill SF3B1-mutant cells, although it alters splicing in both mutant and wild-type backgrounds.35 H3B-8800 increased cell surface DR4 and DR5 expression in SUDHL4, Raji, and Ri-1 cells and synergized with recombinant TRAIL to induce apoptosis, phenocopying JNJ-64619178 (Figure 5H-I; supplemental Figure 7A-D). Similarly, E7070 upregulated cell surface DR4 and/or DR5 expression and synergized with TRAIL in all 3 cell lines (Figure 5J-K; supplemental Figure 7E-H). Unlike JNJ-64619178, both H3B-8800 and E7070 increased the long-to-short DR5 isoform ratio (supplemental Figure 7I-J), indicating altered TNFRSF10B splicing. H3B-8800 reduced the total level of DR4 protein, whereas E7070 induced subtle and variable changes in total DR4 across the 3 cell lines (supplemental Figure 7I-J). These differences suggest that the 3 agents differentially affect TNFRSF10A and TNFRSF10B expression. Nevertheless, all 3 agents increased cell surface DR4 and/or DR5 expression and sensitized cells to TRAIL, indicating a shared downstream response to global splicing disruption. DR4 and DR5 are both functional TRAIL receptors in B-NHL We next evaluated the mechanisms underlying JNJ-64619178- and TRAIL-induced cell killing. To evaluate whether the synergy between JNJ-64619178 and TRAIL requires DR4 or DR5, we tested agonist antibodies targeting DR4 (mapatumumab) or DR5 (conatumumab) in combination with JNJ-64619178 in 6 B-NHL cell lines, including DLBCL (OCI-Ly1, SUDHL4, and TMD8), MCL (Mino), and BL (Raji and BL-70). OCI-Ly1 was the most sensitive to conatumumab, followed by Mino and SUDHL4, and conatumumab synergized with JNJ-64619178 to induce apoptosis in these 3 cell lines (Figure 6A-C). The other 3 cell lines did not respond to conatumumab, and the addition of conatumumab to JNJ-64619178 did not provide further cell killing (data are not shown).Figure 6.JNJ-64619178 synergizes with DR5 agonist antibody conatumumab in vitro. (A-C) OCI-Ly1, SUDHL4, and Mino cells were treated with DMSO or 1μM JNJ-64619178 for 5 days, and conatumumab was added on the last 2 days at indicated dosages in the presence of 1μg/mL soluble protein G. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. (D) Western blot analysis of indicated proteins in OCI-Ly1 expressing control short hairpin RNA (shCTRL) or short hairpin RNA targeting DR5 (shDR5). Densitometry of the DR5 long band and DR5 short band was quantified and normalized to that of GAPDH first, then normalized to the respective shCTRL condition. (E) OCI-Ly1, expressing shCTRL or shDR5, were treated with the combination of JNJ-64619178 and conatumumab (similar to panel A); cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Interaction P value was calculated using 2-way analysis of variance. Data are mean ± SD, n = 3. (F) OCI-Ly1 expressing shCTRL or shDR5 were treated with recombinant human TRAIL for 2 days; cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. DR5-L, DR5 long band; DR5-S, DR5 short band; GADPH, glyceraldehyde-3-phosphate dehydrogenase. JNJ-64619178 synergizes with DR5 agonist antibody conatumumab in vitro. (A-C) OCI-Ly1, SUDHL4, and Mino cells were treated with DMSO or 1μM JNJ-64619178 for 5 days, and conatumumab was added on the last 2 days at indicated dosages in the presence of 1μg/mL soluble protein G. Cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. Mean Loewe synergy score was shown in brackets. (D) Western blot analysis of indicated proteins in OCI-Ly1 expressing control short hairpin RNA (shCTRL) or short hairpin RNA targeting DR5 (shDR5). Densitometry of the DR5 long band and DR5 short band was quantified and normalized to that of GAPDH first, then normalized to the respective shCTRL condition. (E) OCI-Ly1, expressing shCTRL or shDR5, were treated with the combination of JNJ-64619178 and conatumumab (similar to panel A); cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Interaction P value was calculated using 2-way analysis of variance. Data are mean ± SD, n = 3. (F) OCI-Ly1 expressing shCTRL or shDR5 were treated with recombinant human TRAIL for 2 days; cell viability was assessed by PI and annexin V staining and normalized to DMSO treatment. Data are presented as mean ± SD, n = 3. DR5-L, DR5 long band; DR5-S, DR5 short band; GADPH, glyceraldehyde-3-phosphate dehydrogenase. We next asked whether DR5 is required to initiate TRAIL-induced apoptosis in OCI-Ly1. We knocked down DR5 in OCI-Ly1 cells (Figure 6D) and found that this significantly protected OCI-Ly1 cells from conatumumab and its combination with JNJ-64619178 (Figure 6E), confirming on-target DR5 involvement. However, DR5 knockdown did not alter sensitivity to recombinant TRAIL (Figure 6F), suggesting that DR4 also contributed to TRAIL-induced cell death in OCI-Ly1. None of these cell lines responded to mapatumumab, regardless of JNJ-64619178 treatment (data are not shown). Interestingly, TMD8, which were insensitive to conatumumab or mapatumumab monotherapy, responded to the combination treatment, and the addition of JNJ-64619178 further enhanced apoptosis (supplemental Figure 8A), suggesting that simultaneous DR4 and DR5 activation is required to initiate extrinsic apoptosis in this cell line. We also observed synergy between PRMT5 inhibition using GSK3326595 and either recombinant TRAIL or conatumumab in OCI-Ly1 cells (supplemental Figure 8B-C). Together, these data suggest heterogeneous reliance on DR5 and DR4 across B-NHL cell lines and indicate that coactivation of both receptors may be necessary for maximal synergy with JNJ-64619178. Combination of JNJ-64619178 and TRAIL receptor agonist suppresses tumor growth in vivo in a DLBCL cell line–derived xenograft mouse model Building upon these observations, we next asked whether JNJ-64619178–mediated sensitization to extrinsic apoptosis could be recapitulated in vivo. Because OCI-Ly1 cells are highly sensitive to DR5 activation, we evaluated the antitumor efficacy of conatumumab in combination with JNJ-64619178 in vivo using the luciferase-expressing OCI-Ly1 cell line–derived xenograft model. After engraftment, mice with comparable tumor burden were randomly assigned into 4 groups: 1) vehicle control; 2) JNJ-64619178; 3) conatumumab; and 4) combination treatment cohorts. Both drugs were well tolerated at the administered doses. Control mice showed rapid tumor progression and developed paralysis within 3 weeks. Similar to the previous in vivo experiments with JNJ-64619178 and venetoclax, JNJ-64619178 and conatumumab each produced only transient tumor control (Figure 7A-B), resulting in improved survival relative to controls (Figure 7C). In contrast, the combination of JNJ-64619178 and conatumumab significantly suppressed tumor growth and prolonged overall survival compared with either single agent alone (Figure 7A-C).Figure 7.JNJ-64619178 synergizes with DR5 agonist antibody conatumumab in vivo in a DLBCL cell line–derived xenograft mouse model. NSG mice transplanted with luciferase-expressing OCI-Ly1 were assigned into 4 groups to receive either vehicle control (n = 5), conatumumab (n = 6), JNJ-64619178 (n = 6), or conatumumab + JNJ-64619178 (n = 6) treatment. (A) Representative BLI from each group. The first BLI was performed on day 0, which is 5 days after injection of tumor cells. Treatment started on day 1. (B) Quantification of BLI signal over time. Graph shows mean ± standard error of the mean. P value was calculated using Student t test by comparing the combination group with other groups. (C) Kaplan-Meier survival curve showing the percentage of mice without symptoms of disease (eg, weight loss, paralysis). P value was calculated using the log-rank test. JNJ-64619178 synergizes with DR5 agonist antibody conatumumab in vivo in a DLBCL cell line–derived xenograft mouse model. NSG mice transplanted with luciferase-expressing OCI-Ly1 were assigned into 4 groups to receive either vehicle control (n = 5), conatumumab (n = 6), JNJ-64619178 (n = 6), or conatumumab + JNJ-64619178 (n = 6) treatment. (A) Representative BLI from each group. The first BLI was performed on day 0, which is 5 days after injection of tumor cells. Treatment started on day 1. (B) Quantification of BLI signal over time. Graph shows mean ± standard error of the mean. P value was calculated using Student t test by comparing the combination group with other groups. (C) Kaplan-Meier survival curve showing the percentage of mice without symptoms of disease (eg, weight loss, paralysis). P value was calculated using the log-rank test. Taken together, these data demonstrate that JNJ-64619178 sensitizes B-NHL cells to extrinsic apoptosis inducers in vivo. Discussion Although PRMT5 promotes the proliferation of malignant B cells and its inhibition delays tumor growth in xenograft models,6,7,10 a clinical trial of the PRMT5 inhibitor JNJ-64619178 demonstrated minimal clinical activity despite robust target engagement.11 To address this discrepancy, we investigated whether PRMT5 inhibition sensitizes malignant B cells to apoptosis, an effect that could be leveraged through rational combinations. Using DBP, we found that PRMT5 inhibition via JNJ-64619178 increased mitochondrial apoptotic priming across B-NHL cell lines and enhanced their sensitivity to BH3 mimetics. JNJ-64619178 also increased cell membrane-bound DR4 and DR5 expression in B-NHL cell lines and enhanced sensitivity to TRAIL-induced extrinsic apoptosis. A phase 1 study of venetoclax in relapsed or refractory B-NHL, including 28 patients with MCL and 34 patients with DLBCL, showed some single-agent activity but generally short-lived responses.13 Consistent with this, our BH3 profiling indicated partial BCL-2 dependence in most DLBCL and MCL cell lines, but venetoclax triggered limited apoptosis because of codependence on other antiapoptotic proteins such as MCL-1, BCL-xL, and BFL-1.29 Thus, drugs that increase the dependence of tumor cells on BCL-2 or reduce their dependence on alternative antiapoptotic proteins may improve venetoclax responses. Our data suggest that PRMT5 inhibition with JNJ-64619178 induces early apoptotic changes that sensitize DLBCL and MCL cells to venetoclax. Another study using the PRMT5 inhibitor, PRT382, also reported synergy with venetoclax in 6 of 9 MCL cell lines.36 To explain resistance to this combination in the remaining 3 cell lines, the authors evaluated synergy-predicting biomarkers and reported a correlation between basal BCL-2 expression and synergy. However, 1 nonsynergistic cell line with the highest BCL-2 expression was excluded as an outlier, whereas 2 others with minimal or no BCL-2 expression were included in the analysis.36 In our study with DLBCL, MCL, and BL cell lines, the absence of BCL-2 protein expression suggested a resistance to venetoclax. However, BCL-2 expression alone did not reliably indicate functional dependence, as observed in Raji and OCI-Ly1R cells. We also found that JNJ-64619178 increased venetoclax sensitivity by enhancing preexisting BCL-2 dependence rather than reprogramming less BCL-2–dependent cells into BCL-2–dependent ones. These findings argue against the use of BCL-2 expression as a predictive biomarker and instead support DBP to define the antiapoptotic dependencies of cancer cells. Thus, PRMT5 inhibitor–induced changes in apoptotic priming may guide selection of the most effective BH3 mimetic for combination therapy. Of note, the study also reported FOXO1-dependent upregulation of BAX expression in some, but not all, MCL cell lines after 6 days of PRT382 treatment. In our study, treatment with JNJ-64619178 for 3 days increased mitochondrial apoptotic priming across DLBCL, MCL, and BL cell lines without inducing BAX upregulation (supplemental Figure 2B). Extending JNJ-64619178 treatment to 5 days similarly failed to increase BAX expression (supplemental Figure 2C), including in the Mino cell line, which did not show BAX upregulation in response to PRT382.36 These findings suggest that increased BAX expression may be context dependent and is not required for the enhanced apoptotic priming observed after JNJ-64619178 treatment. Normal cells resist TRAIL-induced apoptosis through expression of the decoy receptors TRAIL-R3 and TRAIL-R4, and early TRAIL analogs showed limited clinical success.17,18 With better understanding of TRAIL biology and advancements in drug design, more stable, higher-order receptor-clustering agonists have been developed.37,38 ABBV-621 is a hexavalent TRAIL fusion protein that selectively binds TRAIL receptors with nanomolar affinity,37 and was investigated in a phase I study, but is no longer being developed because of limited clinical efficacy.39 INBRX-109, a tetravalent DR5 agonistic antibody, has shown antitumor activity in patients with chondrosarcoma and is in ongoing clinical development.38 In our study, transcriptomic analysis revealed variable expression of DR4, DR5, and decoy receptor 2 (supplemental Figure 6A). These findings suggest that the baseline DR4/DR5 expression ratio, together with decoy receptor expression, may influence responses to DR4- and DR5-targeting agonists. Our study shows that JNJ-64619178 enhances TRAIL responses by upregulating DR4 and DR5 membrane presence. Unlike the heterogeneous responses to venetoclax observed across B-NHL cell lines, JNJ-64619178 consistently enhanced TRAIL-induced apoptosis across multiple cell lines. We also observed that some cell lines require activation of both DR4 and DR5, potentially via DR4/DR5 heterotrimer formation, to initiate apoptosis, indicating that dual-receptor agonists may pair best with JNJ-64619178 across diverse B-cell malignancies. Immune effector cells such as natural killer cells and cytotoxic T lymphocytes typically express low baseline levels of TRAIL; however, TRAIL expression can be induced after cytokine stimulation (eg, interferon beta, interferon gamma, or interleukin-12), enabling engagement of the TRAIL death receptors DR4 and DR5 on target cells and triggering apoptotic signaling.40, 41, 42 Therefore, future studies should investigate whether activated CAR natural killer cells or CAR T cells synergize with JNJ-64619178 to induce extrinsic apoptosis of lymphoma cells. Although previous studies linked splicing modulators to enhanced BH3 mimetic responses,43, 44, 45 the splicing modulators H3B-8800 and E7070 did not increase mitochondrial apoptotic priming in our B-NHL cell lines (data are not shown). Instead, the upregulated cell surface DR4/DR5 expression was observed with all 3 mechanistically distinct splicing modulators, despite differing effects on total protein expression, suggesting that TRAIL receptor upregulation may reflect increased membrane trafficking as a general response to global splicing-alteration stress. Our study has several limitations. Although we observed increased overall mitochondrial apoptotic priming in B-NHL cell lines after JNJ-64619178 treatment, the underlying mechanisms remain unclear. Arginine methylation may act through posttranslational modification to contribute to functional apoptosis resistance, similar to BCL-2 family protein phosphorylation events.46, 47, 48, 49 Moreover, the mechanisms responsible for JNJ-64619178-induced upregulation of cell surface DR4 and DR5 remain unclear. Our data indicate that this effect is independent of changes in total protein expression and may be associated with global alterations in RNA splicing; however, the signaling pathways linking splicing modulation to enhanced death receptor surface presentation require further investigation. Functional annotation of the splicing targets of JNJ-64619178 revealed significant enrichment of genes encoding proteins localized to the endosome and Golgi apparatus (supplemental Figure 9). Both organelles play important roles in regulating protein trafficking, suggesting that altered splicing of components involved in protein trafficking may contribute to enhanced death receptor surface presentation. In addition, although we examined a diverse panel of B-NHL cell lines, validation in primary patient samples and patient-derived xenograft models will be helpful to support the translational relevance of our findings. In conclusion, our study demonstrates that JNJ-64619178 sensitizes B-NHL cells to both intrinsic and extrinsic apoptosis, suggesting novel therapeutic combination approaches for evaluation in patients. Conflict-of-interest disclosure: M.S.D. has received consulting fees from AbbVie, Ascentage Pharma, AstraZeneca, BeOne Medicines, Bristol Myers Squibb, Eli Lilly, Galapagos, Genentech, Janssen, Merck, Nuvalent, and Schrödinger; and research support from Ascentage Pharma, AstraZeneca, MEI Pharma, and Novartis. C.E.R. has received consulting fees from AstraZeneca, BeOne Medicines, Genentech, and Incyte; and has received institutional research funding from BeOne Medicines, Eli Lilly, and Genentech. The remaining authors declare no competing financial interests.