YMN-V115: a novel humanized BCMA/GPRC5D/CD3 trispecific antibody in relapsed/refractory multiple myeloma

Abstract

Background

Multiple myeloma (MM) is a hematologic malignancy characterized by a high relapse rate despite recent therapeutic advances. B-cell maturation antigen (BCMA) and G protein-coupled receptor class C group 5 member D (GPRC5D) are two well-validated targets, with Food and Drug Administration-approved bispecific antibodies demonstrating clinical efficacy; however, their effectiveness is often limited by antigen escape and immune evasion.

Methods

This study presents a novel humanized trispecific antibody, YMN-V115, designed in a 1+1+1 format with one binding site each for BCMA, GPRC5D, and CD3. This structural configuration was strategically selected to maximize tumor-specific T cell engagement through dual-antigen targeting while minimizing the risk of CD3-mediated off-tumor toxicity.

Results

YMN-V115 exhibited superior cytotoxic activity against MM cells in both in vitro and in vivo models, outperforming its bispecific counterparts. Notably, its efficacy was maintained under soluble BCMA-saturated conditions, demonstrating resistance to decoy inhibition. Moreover, YMN-V115 effectively addressed antigenic heterogeneity by targeting both BCMA⁺ and GPRC5D⁺ tumor subsets, resulting in enhanced T-cell activation and durable tumor control.

Conclusions

YMN-V115 represents a next-generation trispecific T-cell engager that effectively addresses antigenic heterogeneity and immune evasion in MM. These preclinical findings support its further clinical development as a promising therapeutic strategy for relapsed or refractory MM.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• B-cell maturation antigen (BCMA)-directed and G protein-coupled receptor class C group 5 member D (GPRC5D)-directed T cell–engaging bispecific antibodies induce meaningful responses in multiple myeloma. However, antigen heterogeneity, antigen loss, and soluble BCMA-mediated resistance limit the durability of clinical benefit.

WHAT THIS STUDY ADDS

• This study describes YMN-V115, a humanized BCMA×GPRC5D×CD3 trispecific T cell engager with a 1+1+1 binding format. YMN-V115 demonstrated superior antimyeloma activity compared with bispecific antibodies across models with heterogeneous antigen expression and retained activity in the presence of soluble BCMA.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These data support dual-tumor antigen targeting within a single trispecific molecule as an approach to mitigate antigen escape in multiple myeloma. YMN-V115 warrants further clinical evaluation as a next-generation T cell–redirecting therapy for relapsed or refractory disease.

Introduction

Multiple myeloma (MM) is a hematologic malignancy characterized by the clonal proliferation of plasma cells in the bone marrow that leads to the overproduction of monoclonal immunoglobulin, bone lesions, and hypercalcemia.1,3 MM is the second most prevalent hematologic cancer globally and is associated with a relatively poor prognosis compared with many other hematologic malignancies.¹ In relapsed/refractory MM (RRMM), precursor cells that are double- or triple-negative for key targets such as G protein-coupled receptor class C group 5 member D (GPRC5D), B-cell maturation antigen (BCMA), and FcRH5 are relatively rare. Nevertheless, heterogeneous expression of these antigens across myeloma clones contributes to therapeutic resistance. Notably, BCMA and GPRC5D are coexpressed in nearly all MM cases, although their expression levels may differ between subpopulations.⁴ This highlights the potential benefit of dual-targeting therapeutic strategies that engage both BCMA and GPRC5D to overcome tumor heterogeneity.

To date, four bispecific antibodies (bsAbs) have been approved by the US Food and Drug Administration for the treatment of RRMM: BCMA × CD3 elranatamab (Elrexfio; PF-06863135/PF-3135, Pfizer),⁵ BCMA × CD3 teclistamab (Tecvayli; JNJ-64007957, Johnson & Johnson),⁶ GPRC5D × CD3 talquetamab (Talvey; JNJ-64407564, Johnson & Johnson),⁷ and BCMA × CD3 linvoseltamab (Regeneron).⁸ Although T-cell engagers (TCEs) targeting BCMA or GPRC5D have demonstrated substantial clinical activity, a substantial proportion of patients still fail to respond. In heavily pretreated patients with prior exposure to anti-BCMA therapy, the MajesTEC-1 and MagnetisMM-1 trials showed comparable efficacy for teclistamab and elranatamab, yet approximately 40% of patients did not achieve a clinical response.⁹ In comparison, in the MonumenTAL-1 phase 1/2 study (NCT03399799) of the GPRC5D-directed TCE talquetamab, 26–27% of patients failed to reach an objective response.⁹ Moreover, in a follow-up study by Holly Lee et al, among 24 patients previously treated with anti-BCMA or anti-GPRC5D TCEs, 16 experienced relapse, including 9 relapses occurring within 6 months, underscoring the persistent challenges associated with the durability of single-antigen TCE therapies.¹⁰ Key resistance mechanisms include high disease burden and soluble BCMA (sBCMA)-mediated target sink, T-cell exhaustion or dysfunction, and antigen escape.⁹ Therefore, next-generation strategies such as multi-antigen targeting are being explored to mitigate clonal escape and eliminate antigen-low reservoirs before overt relapse. To this end, trispecific antibodies (TriTEs) have emerged as a promising class of therapeutics. Recent clinical studies have evaluated the combination of bispecifics (eg, teclistamab+talquetamab), with results from the 2023 European Hematology Association (EHA) Congress showing an overall response rate (ORR) of 86.6% and ≥complete response (CR) of 34% in RRMM.¹¹ These findings support the rationale for integrated trispecific designs. Several trispecific antibodies targeting BCMA/GPRC5D/CD3 are currently under clinical development, including JNJ-79635322, IBI3003, SIM0500, and MBS314, some of which have shown promising preclinical activity or received Fast Track designation.12,16

In this study, we developed a structurally optimized trispecific T-cell engager (TriTE), YMN-V115. This novel TriTE integrates a high-affinity nanobody (Nb) and a high-affinity single-chain variable fragment (scFv) to simultaneously target BCMA, GPRC5D, and CD3 (figure 1). YMN-V115 exhibited favorable stability and high expression levels in vitro, along with potent picomolar-range cytotoxicity against BCMA⁺ and/or GPRC5D⁺ MM cells, outperforming both bsAbs and other trispecific constructs. Importantly, its activity was not affected by the presence of sBCMA. In vivo, YMN-V115 induced marked tumor regression and prolonged survival in both cell line-derived xenograft (CDX) and patient-derived xenograft (PDX) models. These findings underscore the therapeutic potential of YMN-V115 in overcoming antigenic heterogeneity and immune escape in RRMM.

Figure 1. Antigen heterogeneity and dual-targeting immunotherapeutic approaches in MM. YMN-V115 is a humanized BCMA×GPRC5D × CD3 trispecific T-cell engager with a dual-antigen binding geometry. YMN-V115 redirects T cells to target BCMA⁺ and GPRC5D⁺ MM cells, thereby addressing antigen heterogeneity and relapse-associated antigen downregulation, while retaining activity in the presence of sBCMA. (A) Antigen heterogeneity in MM and limitations of current therapies; (B) Design of the YMN-V115 trispecific antibody; (C) Mechanism of action and functional outcomes. BCMA, B-cell maturation antigen; GPRC5D, G protein-coupled receptor class C group 5 member D; MM, multiple myeloma; Nb, nanobody; sBCMA, soluble BCMA; scFv, single-chain variable fragment.

Results

Independent expression of GPRC5D and BCMA in multiple myeloma

To assess the relationship between BCMA and GPRC5D expression in MM, transcriptomic data from 600 newly diagnosed patients with MM were analyzed using the Multiple Myeloma Research Foundation CoMMpass dataset (GSE39754). Nearly all samples expressed either BCMA, GPRC5D, or both, although at varying levels (figure 2A). Correlation analysis revealed a weak positive correlation (R=0.36) between the two targets (figure 2B,C), suggesting largely independent expression patterns. These findings highlight GPRC5D as a complementary therapeutic target to BCMA, particularly for overcoming resistance in BCMA-low or BCMA-negative disease subclones.

Figure 2. The expression of GPRC5D on r/r MM and generation of anti-GPRC5D clone 48# mAb. (A) The transcript expression levels of GPRC5D and BCMA in patients with MM (n=600) from MMRF-CoMMpass dataset. (B–C). Correlation analysis showing a low association between GPRC5D and BCMA expression (r=0.36). (D) Serum titers against GPRC5D in immunized mice (A90 and A91) measured 1 week after the final boost; recombinant B7H3-His served as a negative control. (E) Indirect ELISA demonstrating the binding activity of anti-GPRC5D antibodies to recombinant human GPRC5D protein. (F) Immunofluorescence microscopy (100×) confirming strong binding of anti-GPRC5D mAbs to GPRC5D-GFP-HeLa cells. Scale bars, 5 µm. (G) Flow cytometric analysis of antibody binding to MM cell lines (RPMI 8226, MM.1S) and negative control HEK293 cells. (H) Real-time cell analysis showing potent, GPRC5D-specific cytotoxicity of anti-GPRC5D mAbs. BCMA, B-cell maturation antigen; GPRC5D, G protein-coupled receptor class C group 5 member D; mAb, monoclonal antibody; MM, multiple myeloma; MMRF, Multiple Myeloma Research Foundation; r/r MM, relapsed/refractory multiple myeloma.

Generation, characterization, and humanization of an antihuman GPRC5D monoclonal antibody

Our team has developed anti-BCMA (clone, hu388) Nb in a previous study.¹⁷ In this study, we developed a novel antibody targeting GPRC5D. To generate the immunogen, recombinant human GPRC5D protein with a C-terminal polyhistidine tag was expressed in human HEK293 cells and subsequently purified (online supplemental figure S1A). The bioactivity of the purified protein was confirmed using the GPRC5D antibody, as shown in online supplemental figure S1B. After multiple rounds of immunization, serum titers reached 1:256 000(figure 2D). Hybridoma screening was then performed using GPRC5D-positive MM.1S cells and GPRC5D-negative MOLM-13 cells, leading to the identification of four monoclonal antibodies with strong binding affinity (online supplemental figure S1C,D). Sequencing was performed, and human GPRC5D scFv-hFc chimeric antibodies were expressed in human embryonic kidney 293T cells. The recombinant chimeric antibodies were purified using protein A affinity chromatography (online supplemental figure S1E). Clone 48# exhibited the highest binding activity, as confirmed by ELISA, immunofluorescence, and flow cytometry (figure 2E–G). Using real-time cell analysis (RTCA), we observed that Clone 48# induced potent and specific lysis of GPRC5D⁺ HeLa cells when co-cultured with T cells at an effector-to-target (E:T) ratio of 2:1 in the presence of 10 pM GPRC5D×CD3 bsAb (figure 2H).

To minimize human anti-mouse antibody responses, clone 48# was humanized using two strategies: (1) deep learning-based optimization (hu481) via the BioPhi platform (https://biophi.dichlab.org) and (2) conventional complementarity-determining region (CDR) grafting (hu482). Both variants were expressed and purified to >90% purity (figure 3A). Flow cytometry showed that hu482 retained comparable binding to the parental murine antibody across multiple MM cell lines (RPMI 8226, MM.1S) but not to GPRC5D-knockout cells (figure 3B). Confocal microscopy confirmed specific binding of hu482 to endogenous GPRC5D (figure 3C, online supplemental figure S2C,D). Indirect ELISA indicated that hu481 exhibited a ∼16-fold fold reduction in affinity, whereas hu482 maintained parental-level binding (figure 3D).

Figure 3. Humanization and characterization of the anti-GPRC5D (clone 48#) mAb. (A) SDS-PAGE analysis confirmed the high purity of the humanized anti-GPRC5D scFv-hFc fusion proteins hu481-hFc and hu482-hFc. (B) Flow cytometric assessment of antibody binding to MM cell lines (RPMI 8226, MM.1S) and their GPRC5D knockout variants. (C) Immunofluorescence microscopy (100×) showing specific binding of hu482 to endogenous GPRC5D in MM.1S cells. Scale bar, 5 µm. (D) Indirect ELISA demonstrating comparable binding affinity of hu482 and the parental mu48 antibody. (E) Cross-species reactivity of hu482 analyzed by flow cytometry using HEK293T cells expressing human, cynomolgus, or murine GPRC5D. (F) Structural docking model of hu482 with GPRC5D predicted by ClusPro and visualized in PyMOL. (G) Schematic representation of GPRC5D chimeras highlighting the NTD, ECLs, TM, and CTD. (H) Flow cytometric evaluation of hu482 binding to HEK293T cells transfected with wild-type mGPRC5D or hGPRC5D mutant constructs. Representative results from two independent experiments are shown. CTD, C-terminal domain; DAPI, (4',6-Diamidino-2-phenylindole) dihydrochloride; ECLs, extracellular loops; GPRC5D, G protein-coupled receptor class C group 5 member D; mAb, monoclonal antibody; MM, multiple myeloma; NTD, N-terminal domain; scFv, single-chain variable fragment; SDS-PAGE, Sodium Dodecyl Sulfate PolyAcrylamide Gel Electrophoresis; TM, transmembrane.

Cross-species reactivity analysis demonstrated that hu482 strongly recognized cynomolgus monkey GPRC5D but not murine GPRC5D (figure 3E). To define the binding epitope of hu482, we used AlphaFold3 to predict the structures of hu482 (scFv) (online supplemental figure S2A) and human GPRC5D (UniProt accession Q9NZD1) (online supplemental figure S2B). Structural modeling using AlphaFold3,¹⁸ followed by protein-protein docking using ClusPro,¹⁹ suggested that hu482 primarily engages the extracellular loop 1 (hECL1) of GPRC5D, with additional contacts involving the N-terminal domain (hNTD) and hECL3 (figure 3F–H). Collectively, these results identify hu482 as a high-affinity, humanized anti-GPRC5D antibody suitable for further functional development.

Molecule design and characterization of YMN-V115 trispecific antibody

To optimize T-cell redirection against MM, we engineered two trispecific T-cell engagers (TriTEs), YMN-V115 and YMN-V118 (figure 4A). YMN-V115 is a 1+1+1 asymmetric trispecific antibody incorporating an anti-BCMA Nb previously developed and published by our group (clone hu388),¹⁷ an anti-GPRC5D scFv generated in this study via hybridoma screening (clone hu482), and an anti-CD3 scFv based on the OKT3 variable-region sequence (clone OKT3-scFv).²⁰ The Fc region contains knobs-into-holes mutations to ensure correct heterodimer assembly and prevent homodimer formation. YMN-V118 adopts a 2+1+1 configuration and contains an additional copy of the anti-GPRC5D scFv (clone hu482) to enhance avidity. Schematic diagrams of the plasmid constructs for YMN-V115 and YMN-V118 are shown in online supplemental figure S3A.

Figure 4. Structural design and characterization of BCMA/GPRC5D/CD3 TriTEs. (A) Schematic representation of YMN-V115 and YMN-V118, trispecific antibodies targeting BCMA (green, clone:hu388), GPRC5D (blue, clone:hu482), and CD3 (orange, clone: OKT3) based on a knob-into-hole platform. (B) Size-exclusion chromatography analysis showing high purity with minimal aggregation. (C–D) Binding activity of YMN-V115 and YMN-V118 compared with isotype control in MM.1R (C) and RPMI 8226 (D) cell panels, including wild-type, BCMA-KO, GPRC5D-KO, and double-KO variants, confirming target-specific engagement of each arm. (E–G) Thermal stability of YMN-V115 assessed by Tm, Tagg, and DLS using the UNcle system (Unchained Labs). Data represent two independent experiments. (H) YMN-V115 maintained stable binding to MM.1R cells across serial dilutions of human serum. (I) sBCMA did not interfere with YMN-V115 binding to MM.1R cells. (J) Binding affinity remained unchanged in the presence of patient serum samples (n=3). APC, Allophycocyanin; BCMA, B-cell maturation antigen; DLS, dynamic light scattering; GPRC5D, G protein-coupled receptor class C group 5 member D; MFI, mean fluorescence intensity; MM, multiple myeloma; mAU, milli–absorbance units; sBCMA, soluble BCMA; Tagg, aggregation onset temperature; Tm, melting temperature; TriTE, trispecific T-cell engager.

Both TriTEs were expressed and purified to >95% purity, as verified by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and size-exclusion chromatography (SEC) (figure 4B, online supplemental figure S3B,C). YMN-V115 demonstrated robust and reproducible expression across three independent production batches, with an average yield of 766 mg/L (online supplemental figure S3D), indicating excellent manufacturability.

Surface expression levels of BCMA and GPRC5D were quantified using the QuantiBRITE PE system across MM cell lines (RPMI 8226, MM.1S, MM.1R) (online supplemental table S1). Flow cytometric analysis revealed that YMN-V115 bound to both antigens with high affinity, exhibiting EC₅₀ values of ∼2.4 nM and ∼4.5 nM for wild-type MM.1R and RPMI 8226 cells, respectively (figure 4C,D, online supplemental tables S2 and S3). Selective monovalent binding was confirmed in BCMA-knockout or GPRC5D-knockout cell lines. Binding of the CD3 arm to primary T cells from two healthy donors yielded EC₅₀ values of 31.2 nM and 8.4 nM, respectively (online supplemental figure S3E, online supplemental table S4). Importantly, YMN-V115 showed minimal off-target reactivity to normal cell types, including HUVEC, BEAS-2B, and HEK293T (online supplemental figure S3F), indicating good tumor selectivity.

In addition, antibody stability is a critical prerequisite for ensuring optimal function, as it is essential for maintaining favorable pharmacokinetics and therapeutic efficacy. The stability of YMN-V115 was assessed using the UNCLE Protein Stability Analysis System. The assay employs a temperature ramp with a total run time of approximately 2 hours, during which the melting temperature (Tm) and aggregation onset temperature (Tagg) values are determined. YMN-V115 exhibited a first melting transition (Tm1) of 69.4°C, an Tagg of 66.9°C, and a stable hydrodynamic profile by dynamic light scattering (figure 4E–G), demonstrating its favorable biophysical characteristics. Moreover, YMN-V115 retained full binding capacity to MM.1R cells even in media containing up to 90% human serum (figure 4H), indicating strong serum stability and resistance to non-specific interference.

Resistance of YMN-V115 to soluble BCMA interference

BCMA shedding is considered a major barrier to the efficacy of BCMA-targeted therapies, with elevated serum sBCMA levels associated with disease progression and poor clinical prognosis.²¹ Previous studies have reported that sBCMA concentrations in healthy individuals range from 18.78 to 180.39 ng/mL, whereas patients with relapsed MM exhibit significantly higher levels, ranging from 100 to 1700 ng/mL.^{22 23}

To test whether YMN-V115 is affected by sBCMA, we assessed its binding to MM.1R cells in the presence of increasing concentrations of recombinant BCMA-mFc (0–2000 ng/mL), corresponding to or exceeding clinically observed levels. YMN-V115 binding remained unaffected even at 2000 ng/mL sBCMA (figure 4I), indicating preserved antigen recognition under pathological conditions.

Next, we evaluated the functional impact of sBCMA on YMN-V115-mediated cytotoxicity. Co-culture of BCMA⁺ HeLa cells with activated T cells in the presence of 10 pM YMN-V115 and increasing BCMA-mFc concentrations (0–2 µg/mL) revealed no reduction in cytotoxic activity (online supplemental figure S3G,H). To more closely mimic clinical conditions, serum samples from three patients with MM, each containing elevated levels of endogenous sBCMA (online supplemental figure S3I), were coincubated with normal human serum. Flow cytometric analysis showed that the presence of patient-derived serum did not alter YMN-V115 binding to MM.1R cells (figure 4J).

Collectively, these findings demonstrate that YMN-V115 maintains stable binding, potent effector activity, and functional resistance to sBCMA, supporting its potential to overcome one of the major resistance mechanisms of BCMA-targeted therapies.

YMN-V115 shows enhanced specific cytotoxic responses compared with monotargeted TCE in vitro

The mechanism of T cell-mediated cytotoxicity induced by YMN-V115 is illustrated in figure 5A. To evaluate cytotoxic efficiency, MM.1R-mCherry.ffLuc cells expressing both GPRC5D and BCMA were co-cultured with healthy donor-derived T cells at various E:T ratios in the presence of 10 pM IgG4 Isotype antibody, bsAbs (JNJ-7957 or JNJ-7564), combination treatment groups (JNJ-7957/JNJ-7564, at a ratio of 1:1), or trispecific antibody (TriTEs; YMN-V115, YMN-V118). At a high E:T ratio of 2:1, all antibody treatments except YMN-V118 induced strong cytotoxic responses, exceeding 80%. However, at lower E:T ratios, the cytotoxicity of most treatments declined markedly. Notably, YMN-V115 maintained high killing efficiency, achieving >60% cytotoxicity even at a 0.5:1 E:T ratio, whereas YMN-V118 exhibited minimal activity under all conditions tested (figure 5B). The isotype control did not elicit any cytotoxic effects. Corresponding elevations in tumor necrosis factor-alpha (TNF-α), interferon-gamma (IFN-γ), and interleukin-2 (IL-2) levels were observed and correlated with cytolytic activity. As demonstrated in figure 5C–E, YMN-V115 induced moderate levels of IL-2, TNF-α, and IFN-γ in co-culture assays, confirming its capacity to activate T cells without triggering uncontrolled cytokine release.

Figure 5. YMN-V115 shows enhanced specific cytotoxic responses compared with monotargeted TCE in vitro. (A) A schematic representation illustrates the T-cell-mediated cytotoxicity induced by YMN-V115 against MM cells. (B–F). The study presents a comparative analysis of the cytotoxic effects of YMN-V115 and the corresponding bsAbs at varying E/T ratios, using T cells co-cultured with MM.1R-mCherry.ffLuc target cells over a 24-hour period. (C–E) Cytokine secretion profiles (IL-6, TNF-α, IFN-γ, and IL-2) were measured in co-cultures treated with 10 pM antibodies at E/T ratios of 2:1, 1:1, and 0.5:1. Data represent mean±SD from three independent experiments. Statistical analysis was performed using two-way ANOVA with Tukey’s multiple comparison test (*p<0.05; **p<0.01; ***p<0.001; ns, not significant). Additionally, IgG4 isotype antibody, bsAbs (JNJ-7957 or JNJ-7564), combination treatment groups (JNJ-7957/JNJ-7564, at a ratio of 1:1), or trispecific antibody (TriTEs; YMN-V115, YMN-V118) were tested against BCMA⁺/GPRC5D⁺ cell lines (MM.1R, MM.1S, RPMI8226) (F) and BCMA⁻/GPRC5D⁻ cell lines (MV-411, MOLM-13, U937) (G). ANOVA, analysis of variance; BCMA, B-cell maturation antigen; bsAbs, bispecific antibodies; E:T, effector-to-target; GPRC5D, G protein-coupled receptor class C group 5 member D; IFN-γ, interferon-gamma; IL, interleukin; MM, multiple myeloma; TCE, T-cell engager; TNF-α, tumor necrosis factor-alpha.

RTCA using impedance-based measurements further confirmed the potent cytotoxic activity of YMN-V115. When applied to GPRC5D⁺/BCMA⁺ HeLa cells and parental HeLa controls, YMN-V115 induced a marked reduction in cell viability in the double-positive line, with no measurable cytotoxicity in parental cells (online supplemental figure S4A,B).

To assess broader cytotoxic potential, YMN-V115 was tested across MM cell lines endogenously expressing GPRC5D and BCMA (MM.1R-mCherry.ffLuc, MM.1S-mCherry.ffLuc, RPMI 8226-mCherry.ffLuc; figure 5F), as well as on cell lines lacking both targets (MV-4–11-mCherry.ffLuc, MOLM-13-mCherry.ffLuc, U937-mCherry.ffLuc; figure 5G). YMN-V115 consistently demonstrated superior cytotoxicity in GPRC5D⁺/BCMA⁺ cell lines compared with either bispecific T cell engager (BiTE) monotherapy or their combination. Importantly, no cytotoxic activity was observed in double-negative cell lines, indicating high target specificity. As expected, the isotype control did not elicit any cytotoxic effects.

YMN-V115 showed enhanced in vivo antitumor efficacy

To assess the in vivo activity of YMN-V115 and bsAbs (JNJ-7957 and JNJ-7564), NCG mice were implanted with MM.1S-mCherry.ffLuc cells or RPMI 8226-mCherry cells. Seven days post-tumor cell inoculation, activated T cells were administered intravenously. From day 9 onward, mice received phosphate-buffered saline (PBS), mice received different treatment regimens: PBS control, single bsAbs (JNJ-7957 or JNJ-7564), combination bsAbs (JNJ-7957/JNJ-7564), sequential treatment (bsAb first, followed by YMN-V115 on tumor relapse), or YMN-V115 alone (trispecific antibody, TriTE). All combination treatments were administered at a 1:1 ratio. In the MM.1S model, treatments were given every 2 days for three consecutive doses, and in the RPMI 8226 model, for five doses. In vivo imaging was conducted weekly until tumor relapse, after which the same treatment regimen was reinitiated (figures6A 7A).

Figure 6. YMN-V115 shows superior antitumor activity to TCE benchmarks in the MM.1S xenograft model. (A) Schematic of the experimental procedure. Beginning on day 9, mice received different treatment regimens: PBS control, single bsAb treatment (JNJ-7957 or JNJ-7564), combination bsAb treatment (JNJ-7957/JNJ-7564), sequential treatment (bsAb first, followed by YMN-V115 on tumor relapse), or YMN-V115 alone (trispecific antibody, TriTE). All combination treatments were administered at a 1:1 ratio. (B) Corresponding bioluminescence images of the mice are provided, with color coding indicating luminescence intensity (red indicating the highest and blue the lowest). (C) Average radiance quantification (p/s/cm²/sr) of the luminescence is shown. (D) Survival curves of MM.1S xenograft mice treated with YMN-V115 and control formulations (n=5). (E–G). Serum proinflammatory cytokines in mice. After 1 day of treatment, the secretion levels of TNF-α, IFN-γ, and IL-2 in the MM.1S model. Representation of average fluorescence intensity in the limb (H) and in major organs, including the heart, liver, spleen, lung, kidney, and brain (I) on day 28. Data are presented as mean±SD. Statistical significance was assessed using one-way or two-way ANOVA, with significance levels indicated as follows: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; bsAb, bispecific antibody; IFN-γ, interferon-gamma; IL, interleukin; i.v., intravenous; MM, multiple myeloma; ns, not significant; PBS, phosphate-buffered saline; TCE, T-cell engager; TNF-α, tumor necrosis factor-alpha.

Figure 7. YMN-V115 shows superior antitumor activity to TCE benchmarks in the RPMI 8226 xenograft model of MM. (A) Schematic of the experimental procedure. Beginning on day 9, mice received different treatment regimens: PBS control, single bsAb treatment (JNJ-7957 or JNJ-7564), combination bsAb treatment (JNJ-7957/JNJ-7564), sequential treatment (bsAb first, followed by YMN-V115 on tumor relapse), or YMN-V115 alone (trispecific antibody, TriTE). All combination treatments were administered at a 1:1 ratio. (B) Bioluminescence imaging of the mice was conducted, with color gradients indicating luminescence intensity (red representing the highest and blue the lowest). (C) The quantification of average radiance (p/s/cm²/sr) of the luminescence is depicted. (D) Survival curves of RPMI 8226 xenograft mice treated with YMN-V115 and control formulations. (E–G). Serum proinflammatory cytokines in mice. After 1 day of treatment, the secretion levels of TNF-α, IFN-γ, and IL-2 in the MM.1S model. Representation of average fluorescence intensity in the limb (H) and in major organs, including the heart, liver, spleen, lung, kidney, and brain (I) on day 28. Data are presented as mean±SD. Statistical significance was assessed using one-way or two-way ANOVA, with significance levels indicated as follows: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; bsAb, bispecific antibody; IFN-γ, interferon-gamma; IL, interleukin; i.v., intravenous; MM, multiple myeloma; ns, not significant; PBS, phosphate-buffered saline; TCE, T-cell engager; TNF-α, tumor necrosis factor-alpha.

In both models, tumor recurrence occurred in the bsAbs group on day 28. Flow cytometric analysis of peripheral blood from relapsed mice showed that, in both the RPMI-8226 and MM.1S CDX models, the majority of relapsed tumors exhibited partial downregulation of BCMA and/or GPRC5D expression (online supplemental figures S5A,B and S6A,B), consistent with antigen escape as a contributing factor to tumor relapse. Following retreatment, combination therapy and YMN-V115 significantly inhibited tumor regrowth, while no recurrence was observed in the YMN-V115 group throughout the study (figures6B,C 7B,C). Both combination therapy and YMN-V115 markedly prolonged survival (p<0.0001; figures6D 7D). Serum cytokine analysis 24 hours after the second dose showed elevated IL-2, IFN-γ, and TNF-α levels in the combination and YMN-V115 groups compared with bsAb groups, consistent with their enhanced antitumor activity (figures6EG and 7E).

To assess the ability of YMN-V115 to overcome immune escape, YMN-V115 was administered sequentially after relapse in bsAb-treated mice. Three days after the second dose, major organs were collected for ex vivo bioluminescence imaging (online supplemental figures S5C,D and S6C,D). PBS-treated mice exhibited extensive metastases, whereas bsAb-treated mice showed weak residual signals, indicating limited tumor escape. In contrast, no fluorescent signals were detected in either the combination or YMN-V115 groups, confirming complete suppression of tumor dissemination (figures6H,I 7H,I).

Overall, PBS-treated mice exhibited rapid disease progression, with a median survival of approximately 30 days. Treatment with BsAbs as monotherapy resulted in only a modest extension of survival and failed to achieve complete tumor eradication. Combination treatment with BsAbs induced tumor regression; however, relapse was observed during follow-up, leading to limited survival benefit. Under the initially tested conditions, the antitumor efficacy of YMN-V115 monotherapy was comparable to that of the BsAb-based combination treatments, as assessed by bioluminescence imaging and survival analysis. On extension of the observation period to 80 days, YMN-V115 monotherapy achieved durable tumor clearance without detectable relapse or metastasis and was associated with a statistically significant improvement in overall survival compared with the BsAb combination treatment group. In addition, YMN-V115 remained effective in mice that relapsed following BsAb treatment.

YMN-V115 treatment was well tolerated, with no significant weight loss, behavioral abnormalities, or clinical toxicity. Histopathological examination of major organs revealed no pathological changes or treatment-related mortality (online supplemental figures S5E, S6E), supporting a favorable safety profile at tested doses.

Efficacy of YMN-V115 against primary MM bone marrow cells in a PDX model

Cancer is a polyclonal disease where each clone may express different types of tumor antigens. Multiple clones can be observed within the same patient, and these clones may express tumor antigens at varying levels (or not at all) in dominant clones. This heterogeneity could significantly limit the efficacy of TCEs (bsAbs) that target a single tumor antigen. Therefore, we next examined the cytotoxic effects of YMN-V115 on primary MM patient samples (n=3). Notably, YMN-V115 exhibited enhanced cytotoxicity across all primary MM samples (figure 8A) and effectively stimulated T cell activation (figure 8B). Additionally, cytokine production, including IL-2, IFN-γ, and TNF-α, was significantly elevated at a concentration of 100 pM compared with bsAbs and bsAbs combination (figure 8C–E).

Figure 8. YMN-V115 demonstrates effective elimination of primary MM patient tumor cells both in vitro and in vivo. (A–E) Cytotoxicity comparison of YMN-V115, bsAbs and bsAbs combinations group against 3 primary MM cells after 24 hours of incubation with expanded T cells at an E/T ratio of 1:1. Experiments were performed in triplicate and repeated three times with similar results. (A) YMN-V115 facilitated the cytotoxicity of CD138+cells from the bone marrow of patients with MM and promoted T cell activation (B) (indicated by CD69+expression). (C–E) Inflammatory cytokines (IL-2, IFN-γ, and TNF-α) released from T cells cocultured with primary MM cells in the presence of bsAbs or TriTE (100 pM) for 24 hours at an E/T ratio of 1:1 in triplicate. (F) Schematic representation showing the experimental design. (G) The expression profiles of BCMA and GPRC5D in patient cells. (H–K) In vivo efficacy of BCMA/GPRC5D/CD3 TriTE in MM PDX models. (H, J) Tumor volume was measured over time using calipers in the MM1 (H) and MM2 (J) PDX models. Tumor growth dynamics were assessed by plotting calculated tumor volumes over the treatment period. (I, K) Kaplan-Meier survival curves for the MM1 (I) and MM2 (K) models. Survival differences between treatment groups were analyzed using the log-rank (Mantel-Cox) test. Statistical comparisons were conducted using one-way or two-way ANOVA followed by Tukey’s multiple comparison test. Exact p values are reported; p<0.05 was considered statistically significant. ANOVA, analysis of variance; BCMA, B-cell maturation antigen; bsAbs, bispecific antibodies; E/T, effector-to-target; GPRC5D, G protein-coupled receptor class C group 5 member D; IFN-γ, interferon-gamma; IL, interleukin; MFI, mean fluorescence intensity; MM, multiple myeloma; PBS, phosphate-buffered saline; PDX, patient-derived xenograft; s.c., subcutaneously; TNF-α, tumor necrosis factor-alpha; TriTE, trispecific T-cell engager.

To evaluate the efficacy of YMN-V115 in patients with MM, we selected two primary MM samples (MM1 and MM2) to establish a clinically relevant PDX model (figure 8F). The antigenic expression profiles of BCMA and GPRC5D in the patient’s cells are shown in figure 8G. As depicted in figure 8H–K, combination treatment groups and YMN-V115 group demonstrated sustained antitumor activity and significantly prolonged mouse survival compared with the bsAb groups. These findings suggest that YMN-V115 holds promise as an effective therapeutic option for patients with MM.

Discussion

T cell-redirecting bsAbs targeting BCMA or GPRC5D have led to meaningful clinical responses in patients with RRMM. Nonetheless, sustained disease control remains challenging, largely due to antigen loss, heterogeneous target expression, and the inhibitory effects of sBCMA.^{10 24}

Here, we report the design and preclinical characterization of YMN-V115, a novel BCMA×GPRC5D×CD3 trispecific T cell engager (TriTE), and demonstrate its potential to overcome several of these key resistance mechanisms. BCMA and GPRC5D exhibit largely independent expression patterns in primary MM samples, consistent with prior reports,⁴ providing a strong biological rationale for dual-antigen targeting. Leveraging this complementary expression, YMN-V115 consistently demonstrated superior antitumor activity compared with single BsAbs or their combination across models with varying antigen densities and expression ratios. These results indicate that simultaneous engagement of BCMA and GPRC5D within a single molecule provides functional benefits that extend beyond coadministration of separate agents, particularly in the context of antigen heterogeneity.

The activity of YMN-V115 is closely linked to its molecular design. Unlike other BCMAxGPRC5DxCD3 trispecific candidates currently in clinical development, including JNJ-79635322, IBI3003, SIM0500, and MBS314,14,16 YMN-V115 employs a hybrid design that integrates a high-affinity Nb with a high-affinity scFv. This format was associated with favorable expression yield, thermal stability, and low aggregation propensity,25,27 features that are critical for the development of multispecific biologics. Importantly, our data suggest that binding geometry plays a decisive role in functional performance. Although bivalent formats may theoretically increase apparent affinity, spatial constraints, epitope orientation, and membrane topology can limit the effective simultaneous engagement of both arms with GPRC5D.2028,33 In this context, monovalent targeting of each antigen by YMN-V115 appears to support more efficient simultaneous binding to BCMA and GPRC5D, thereby promoting effective T cell redirection.

YMN-V115 also maintained potent cytotoxic activity in the presence of clinically relevant concentrations of sBCMA, a known mediator of resistance to BCMA-directed therapies.^{22 23} This resistance to sBCMA interference likely reflects reduced reliance on BCMA alone, with GPRC5D serving as an additional anchoring target for T cell activation. Such a feature may be particularly relevant in heavily pretreated patients with RRMM, who often exhibit elevated sBCMA levels.²⁴ Moreover, YMN-V115 effectively induced T cell activation and cytokine release in vitro and achieved durable tumor control with significant survival benefit in vivo, even in models with imbalanced antigen expression or following relapse from single-antigen BsAb therapy. These results are consistent with emerging clinical evidence indicating that concurrent targeting of BCMA and GPRC5D enhances therapeutic efficacy,¹¹ and they further support the concept that a single TriTE molecule may enable more uniform and stable target engagement.

Several limitations warrant consideration. The preclinical models used do not fully capture the complexity of the human bone marrow microenvironment or long-term immune-tumor coevolution.³⁴ Moreover, the selective pressures imposed by dual-antigen targeting and the potential for emergent resistance will need to be evaluated in clinical studies. Although YMN-V115 demonstrated high activity at low concentrations in preliminary comparisons, its relative efficacy and safety compared with other advanced trispecific candidates, such as JNJ-79635322,³⁵ remain to be defined through head-to-head studies.

In summary, YMN-V115 combines rational dual-antigen targeting with an optimized trispecific design to address key limitations of current TCEs. Its developability profile, resistance to sBCMA-mediated inhibition, and activity against antigen-heterogeneous myeloma support further clinical investigation in RRMM.

Methods

Cell culture and reagents

The cell lines RPMI 8226, MM.1S, MM.1R, U937, and MV-4–11 were procured from the American Type Culture Collection (Manassas, USA). MOLM-13 cells were acquired from Procell (CL-0681). MM.1S-mCherry.ffLuc, RPMI 8226-mCherry.ffLuc, GPRC5D⁺-HeLa, BCMA⁺-HeLa, and BCMA⁺/GPRC5D⁺-HeLa cell lines were generated by lentiviral transduction. Knockout cell lines were generated using the CRISPR/Cas9 system. Guide RNAs targeting GPRC5D or BCMA were designed using the CHOPCHOP online platform (http://chopchop.cbu.uib.no) and cloned into the lentiCRISPR v2 vector (cat#52961, Addgene, Watertown, Massachusetts, USA). Lentiviral particles were transduced into MM.1S and RPMI 8226 cells, followed by puromycin selection. One week after transduction, cells were stained with YMN-V115, and GPRC5D/BCMA-negative populations were isolated by fluorescence-activated cell sorting (BD Biosciences). MM cell lines were maintained in RPMI-1640 medium (Gibco, California, USA) supplemented with 10% fetal bovine serum (FBS; Gibco). HeLa cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% FBS and 100 U/mL penicillin-streptomycin.

Cloning, expression and purification of recombinant proteins

The full-length GPRC5D gene was cloned into the pcDNA3.1(+) vector (cat# V79020, Thermo Fisher Scientific, Massachusetts, USA). HEK293T cells were cultured in DMEM supplemented with 10% FBS and seeded in 10 cm dishes to reach 70–80% confluency before transfection. Plasmid DNA (10 µg per dish) was mixed with PEI (3 µL per 1 µg DNA; Polysciences, Warrington, Pennsylvania, USA) in Opti-MEM. 8 hours post-transfection, the culture medium was replaced with fresh FreeStyle 293 (Thermo Fisher Scientific). After 7 days of culture, cells and culture supernatants were harvested. Proteins were purified using Ni-NTA affinity chromatography (Cytiva) or Protein A resin (AT Protein A Diamond Ultra, Bestchrom, China). Protein purity was assessed by SDS-PAGE.

Mouse hybridoma generation and screening

Purified full-length GPRC5D protein (100 µg) was emulsified with an equal volume of Freund’s complete adjuvant and administered intramuscularly to 6-week-old female BALB/c mice. Peripheral blood serum was collected prior to immunization and used as a negative control. Booster injections consisting of GPRC5D protein emulsified in Freund’s incomplete adjuvant were administered every 2 weeks for a total of four boosts. Immune serum was collected 7–10 days after the final boost for antibody titer analysis. For hybridoma generation, mice received an intraperitoneal boost 3 days before spleen harvest. Splenocytes were fused with SP2/0 myeloma cells using polyethylene glycol. Positive hybridoma clones were identified by flow cytometry using both positive and negative selection strategies, yielding four clones with strong antigen-binding activity.

RNA was extracted from positive hybridoma cells using the RNAsimple Total RNA Kit, followed by cDNA synthesis using the PrimeScript RT reagent Kit. VH – Variable region of the heavy chain of an antibody and Variable region of the light chain of an antibody were amplified by nested PCR.

Immunofluorescence

Exponentially growing cells were seeded onto plates and incubated overnight. Cells were fixed with 4% paraformaldehyde and blocked with 5% bovine serum albumin. Primary antibodies were incubated for 1 hour at room temperature, followed by incubation with fluorophore-conjugated secondary antibodies for 1 hour. After PBS washes, samples were mounted with DAPI (4′,6-diamidino-2-phenylindole, a fluorescent dye that binds strongly to DNA)-containing antifade medium and imaged using a confocal microscope. The following antibodies were used for the present study: 594-Goat Anti-Mouse Recombinant Secondary Antibody (H+L) (Proteintech, Wuhan, China), Goat anti-Human IgG Fc Recombinant Secondary Antibody, Alexa Fluor 647 (Invitrogen, USA), Alexa Fluor 594 AffiniPure Goat Anti-Human IgG (H+L) (Jackson, Pennsylvania, USA).

Flow cytometry

A total of 1×10⁶ cells were washed once with PBS and incubated with a human Fc receptor blocker for 30 min at room temperature. Cells were subsequently stained with specific antibodies or purified antibodies for 40 min in the dark, washed, resuspended in PBS, and analyzed by flow cytometry. Isotype controls were included in all experiments. Antibodies used included anti-human CD138 (APC), anti-human CD45 (PerCP), anti-human CD3 (BV421), anti-human CD4 (BV510), anti-human CD8 (FITC), anti-human CD69 (APC), anti-human BCMA (PE) were all purchased from BioLegend (San Diego, California, USA), Anti-GPRC5D Antibody (APC) was purchased from (Sino Biological, China).

Humanization of monoclonal antibody

To reduce immunogenicity associated with murine antibodies, humanization was performed using deep learning-based modeling and CDR grafting. Antibody humanization was conducted using BioPhi, an antibody design and humanization platform (https://biophi.dichlab.org), in combination with natural antibody libraries (Sapiens). Humanized anti-GPRC5D antibodies were subsequently expressed and characterized.

Construction of human-murine GPRC5D chimeric mutants for epitope mapping

The extracellular region of GPRC5D consists of an N-terminal domain and three extracellular loop domains (ECL1–ECL3). To determine the epitope recognized by the anti-GPRC5D antibody, human GPRC5D extracellular domains were sequentially replaced with the corresponding murine domains to generate a panel of human-murine chimeric constructs (see schematic in figure 3G). The chimeric plasmids were transfected into HEK293T cells, and antibody binding to each variant was quantified by flow cytometry to identify the specific extracellular domain required for antibody recognition.

TriTE construction, expression, purification, and SEC characterization

Trispecific T cell engagers (TriTEs) were engineered using a knob-into-hole strategy based on a modified human IgG4 Fc scaffold. Plasmids encoding knob and hole chains for each TriTE were mixed at predefined mass ratios and complexed with PEI (PEI: DNA=3:1). Complexes were used to transiently transfect 293FT cells cultured in FreeStyle serum-free medium. Cells were maintained in shaking flasks for 5–7 days. Supernatants were collected, clarified, concentrated, and purified by Protein A/G affinity chromatography, followed by SEC to obtain highly pure, homogeneous proteins.

Human T cell activation

Peripheral blood mononuclear cells from healthy donors were isolated by density gradient centrifugation. T cells were cultured in X-Vivo medium (Lonza) supplemented with 10% heat-inactivated FBS, IL-7 (10 ng/mL), IL-15 (10 ng/mL), and 1% penicillin-streptomycin. T cells were activated for 72 hours using T Cell TransAct (cat#130-111-160, Miltenyi).

Serum sBCMA quantification

Serum sBCMA levels were measured using the Invitrogen Human BCMA ELISA Kit (Catalog Number: EH41RB) following the manufacturer’s instructions. Serum samples were diluted 1:5 prior to loading, and all measurements were performed in duplicate. Standard curves were generated using the recombinant human sBCMA calibrators provided with the kit.

sBCMA binding-interference assay

To evaluate whether sBCMA interferes with the interaction between YMN-V115 and membrane-bound BCMA, a competitive binding assay was performed. Recombinant human sBCMA at various concentrations was diluted in normal human serum and coincubated with YMN-V115 and MM.1S cells at 37°C for 1 hour. Binding was subsequently analyzed by flow cytometry using a Goat anti-Human IgG Fc Recombinant Secondary Antibody conjugated to Alexa Fluor 647 (Invitrogen, USA). Negative control samples treated with hIgG4 isotype control antibody were included to assess non-specific binding and background signal. The resulting mean fluorescence intensity values were normalized to those obtained from the no-sBCMA control condition.

Functional in vitro assays

Real-time cell analysis experiments

Cytotoxicity assays were measured using label-free RTCA instrument and E-Plate 96 (ACEA Biosciences) according to the manufacturer’s protocol. Briefly, each well of the E-Plate 96 was first loaded with 50 µL of DMEM supplemented with 10% FBS. GPRC5D⁺-HeLa, BCMA⁺-HeLa, BCMA⁺/GPRC5D⁺-HeLa, or parental HeLa cells were then seeded in 100 µL of culture medium and incubated at 37°C until the cell index (CI) exceeded 1.0. Effector cells were subsequently added together with antibodies (IgG4 isotype control, BiTEs JNJ-7957 or JNJ-7564), combination treatments (JNJ-7957 and JNJ-7564 at a 1:1 ratio), or trispecific antibody (YMN-V115 or YMN-V118). Impedance was continuously monitored in real time, and CI-time curves were recorded throughout the assay.

Luciferase-based assays for evaluating cytotoxicity and detecting cytokine levels

After co-culturing BiTEs or TriTEs with T cells derived from healthy donors and luciferase-labeled target cells expressing either GPRC5D⁺/BCMA⁺ (MM.1R-Luc, MM.1S-Luc, RPMI 8226-Luc) or GPRC5D⁻/BCMA⁻ (MV-4–11-Luc, MOLM-13-Luc, U937-Luc) for 24 hours, cytotoxicity was quantified through luminescence measurements. Cytokine levels in the culture supernatants, including IFN-γ (Cat# KHC4021), TNF-α (Cat# KAC1751), and IL-2 (Cat# BMS221-2, Invitrogen), were measured using ELISA kits.

Xenograft mouse models using MM cell lines

All animal experiments were conducted in accordance with the protocol approved by the Institutional Animal Care and Use Committee of West China Hospital, Sichuan University, following international, national, and institutional guidelines for the humane treatment of animals and complying with all relevant legislation. The recipient NCG mice (NOD/ShiLtJGpt-Prkdc ^em26Cd52 IL2rg ^em26Cd22 /Gpt) were female, aged 6–8 weeks, and purchased from GemPharmatech Company (Nanjing, China). MM models were generated using MM.1S-mCherry.ffLuc (high GPRC5D/BCMA) or RPMI 8226-mCherry (low GPRC5D/BCMA). On day 1, mice were injected intravenously with tumor cells. On day 8, activated human T cells were infused. On day 9, mice were randomly assigned to the following treatment groups: PBS control, single bsAb treatment (JNJ-7957 or JNJ-7564), combination bsAb treatment (JNJ-7957+JNJ-7564), sequential treatment (bsAb first, followed by YMN-V115 on tumor relapse), or YMN-V115 alone (trispecific antibody, TriTE). All combination treatments were administered at a 1:1 ratio. Treatments were dosed according to the model: MM.1S mice received 0.25 mg/kg per week for three doses, while RPMI 8226 mice received 0.4 mg/kg per week for five doses. The tumor burden was assessed using Tanon ABL X6 imaging (Shanghai, China) following the intraperitoneal administration of D-luciferin at a dosage of 150 mg/kg (Beyotime).

Serum cytokines (IL-2, IFN-γ, TNF-α) were measured 24 hours after the first dose. For metastasis analysis, organs (heart, liver, spleen, lungs, kidneys, brain), and limbs were collected on days 28, 31, or 35 (model-dependent) for ex vivo bioluminescence imaging.

Development of a multiple myeloma patient-derived xenograft model

NCG mice were implanted subcutaneously with 5×10⁶ primary MM cells (MM1 or MM2) suspended in Matrigel. Mice were monitored weekly for body weight, activity, and tumor progression. On day 18, mice received two doses of expanded T cells, followed by intraperitoneal administration of PBS control, JNJ-7957, JNJ-7564, YMN-V115, BiTEs (JNJ-7957+JNJ-7564, 1:1 ratio), or TriTE (YMN-V115) every other day at 0.5 mg/kg for five doses. Serum cytokine levels were measured 24 hours after the first antibody dose.

Mice were euthanized when tumor volume reached the predetermined endpoint (usually ≥1,500 mm³) or when significant clinical symptoms (eg, weight loss of more than 20%) appeared.

Statistical analysis

Statistical analyses were performed using GraphPad Prism. Data are presented as mean±SD. Student’s t-test, one-way or two-way analysis of variance with Tukey’s post hoc test were used for group comparisons. Survival analysis in the MM models was conducted using Kaplan-Meier estimates, with statistical significance evaluated via the log-rank test.

Experiments included three biological replicates per donor and were performed in duplicate. Significance was defined as *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001; with “ns” indicating no significance.

Data availability statement

No data are available.