Tumor-targeted 4-1BB costimulation enhances immune activation and clinical activity of a CEA-directed T-cell engager in microsatellite-stable colorectal cancer

Tumor-targeted 4-1BB costimulation enhances immune activation and clinical activity of a CEA-directed T-cell engager in microsatellite-stable colorectal cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Tumor-targeted 4-1BB costimulation enhances immune activation and clinical activity of a CEA-directed T-cell engager in microsatellite-stable colorectal cancer Axel Boehnke, Ignacio Melero, Tamara Tanos, Emiliano Calvo, Camilla Qvortrup, and 18 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8628656/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 20 Apr, 2026 Read the published version in Nature Medicine → Version 1 posted You are reading this latest preprint version Abstract Cibisatamab is a carcinoembryonic antigen (CEA)–directed T-cell engaging bispecific antibody that has shown evidence of antitumor activity as a single agent in patients with microsatellite-stable (MSS) metastatic colorectal cancer (mCRC) (ClinicalTrials.gov identifier: NCT02324257). Effective T-cell activation requires coordinated signaling through the TCR–CD3 complex and costimulatory pathways, including 4-1BB, which is induced following antigen recognition. We evaluated the combination of cibisatamab with escalating doses of FAP-4-1BBL, a fibroblast activation protein (FAP)–targeted 4-1BB agonist, in an open-label phase 1b study (BP42675; ClinicalTrials.gov identifier: NCT04826003) in patients with MSS mCRC who had progressed after at least two prior lines of therapy. The combination demonstrated a manageable safety profile and induced robust immune activation, characterized by sustained increases in circulating interferon-γ (IFNγ), soluble CD25 (sCD25), and proliferating/ activated CD8⁺ T-cell subsets, together with expansion of memory compartments. Compared with cibisatamab monotherapy, the combination resulted in greater and more durable induction of IFNγ and sCD25 without any exacerbation of cytokine-release–related toxicity. Treatment was associated with numerically higher overall response and disease control rates than previously reported for cibisatamab alone, with an overall response rate of 17.6% (9/51) and a disease control rate of 50.9% (26/51) across dose levels and schedules. Reductions in serum CEA, early decreases in circulating tumor DNA, and increased soluble CD137 were associated with clinical benefit. Paired tumor biopsies showed greater increases of intratumoral CD8⁺ and CD8⁺Ki67⁺ T-cell infiltration relative to cibisatamab monotherapy. Together, these findings support tumor-localized 4-1BB costimulation as a strategy to enhance the biological and clinical activity of T-cell engagers in MSS mCRC and other non-inflamed solid tumors. Health sciences/Medical research Health sciences/Medical research/Outcomes research Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Immune checkpoint inhibitors have substantially improved outcomes in CRC with mismatch repair deficiency or high microsatellite instability (dMMR/MSI-H) 1 . However, more than 85% of patients present with mismatch repair–proficient, microsatellite-stable (pMMR/MSS) disease 2 , a largely non-inflamed tumor type in which immunotherapy has shown limited efficacy 3 , 4 , resulting in a persistently poor prognosis in the metastatic setting. T-cell receptor–engaging therapeutics can activate cytotoxic T lymphocytes and have demonstrated clinical benefit in hematologic malignancies 5 , 6 and, more recently, in selected solid tumors 7 , 8 . Nonetheless, extending these approaches to poorly immunogenic, non-inflamed tumors such as pMMR/MSS CRC remains challenging. Among inducible costimulatory receptors, 4-1BB (CD137) is upregulated following T-cell activation and promotes proliferation, survival, memory differentiation, and cytotoxic function upon engagement with its ligand or agonistic antibodies 9 – 12 Cibisatamab is a 2:1 T-cell bispecific antibody that bivalently binds a tumor-restricted carcinoembryonic antigen (CEA) epitope exposed on malignant cells after proteolytic cleavage, while engaging CD3ε on T cells. CEA is expressed on more than 80% of colorectal cancers on T cells 13 , 14 . By simultaneously binding CEA and CD3ϵ, cibisatamab induces T-cell activation independently of native T-cell receptor specificity, resulting in lymphocyte-mediated tumor cell killing 14 , 15 . In a prior phase 1 study of cibisatamab monotherapy in CEA-positive solid tumors, preliminary antitumor activity was observed, with confirmed partial responses in 4.0% of evaluable participants and a median duration of response of 6.5 months 16 . We hypothesized that the antitumor activity of a T-cell engager such as cibisatamab could be enhanced by spatially restricted costimulation mediated by a fibroblast activation protein (FAP)–targeted 4-1BB ligand (FAP-4-1BBL). This concept is supported by preclinical coculture models of tumor organoids and FAP-expressing cancer-associated fibroblasts 17 . FAP-4-1BBL enables selective 4-1BB engagement within FAP-expressing tissues, including the tumor microenvironment and tumor-draining lymph nodes 11 , 18 . In a first-in-human dose-escalation study, FAP-4-1BBL monotherapy showed modest clinical activity but demonstrated an acceptable safety profile and evidence of increased intratumoral T-cell infiltration was observed 11 . Here, we report the results of the phase 1b dose-escalation study BP42675 evaluating the combination of cibisatamab and FAP-4-1BBL following obinutuzumab pre-treatment in patients with heavily pretreated metastatic MSS CRC. Obinutuzumab pretreatment to deplete B cells was required because of almost constant induction of anti-drug antibodies (ADAs) by Cibisatamab 19 . We further contextualize these findings in the combination setting by comparison with previously reported cibisatamab monotherapy data and associated pharmacodynamic analyses 16 . Our studies assessed the impact of tumor-localized 4-1BB costimulation on immune activation and clinical activity. Hereafter, RO7122290 is uniformly referred to as FAP-4-1BBL to ensure consistent terminology across the manuscript. Results Participants and Treatment Fifty-two participants with pMMR/MSS mCRC who had progressed after at least two prior lines of therapy were enrolled in this open-label, multicenter, Phase 1b dose-escalation study. Baseline characteristics and demographics are summarized in Supplementary Data Table 1 . The study comprised two parts. Part 1 evaluated safety, pharmacokinetics (PK) and pharmacodynamics following escalation of FAP-4-1BBL with weekly dosing (QW). Doses considered safe in Part 1 were explored in Part 2 with FAP-4-1BBL administered every 3 weeks (Q3W) (Supplementary Figs. 1 and 2). In both parts, cibisatamab was administered Q3W. Up to Part 1 Cohort 2 and Part 2 Cohort 1, cibisatamab was given at 100 mg from Cycle 1. Following a Grade 3 CRS event, the Cycle 1 cibisatamab dose was reduced to 60 mg from Part 1 Cohort 3 and Part 2 Cohort 2 onward, with 100 mg from Cycle 2 (cibisatamab 60→100 mg). Cohort-level dosing details are provided in Baseline Characteristics Supplementary Table 1. FAP-4-1BBL escalation in combination with cibisatamab was discontinued for cibisatamab-related strategic reasons (no safety signal observed) before a maximum tolerated dose (MTD) was defined. Anti-drug antibody management Previous clinical experience with cibisatamab has demonstrated the induction of ADAs, primarily mediated by B-cell activation and differentiation into antibody-secreting plasma cells 16 . To address this, the Phase 1b combination study of cibisatamab plus FAP-4-1BBL (BP42675) incorporated B-cell depletion with obinutuzumab as a pre-treatment, with a planned, repeated administration at six months. This strategy was designed to suppress de novo ADA formation while preserving pre-existing humoral immunity mediated by long-lived plasma cells, consistent with prior clinical observations 16 . This study incorporated B-cell depletion with obinutuzumab as a pre-treatment, with a planned, repeat dosing at six months in patients remaining on treatment, to suppress ADA formation while preserving pre-existing humoral immunity mediated by long-lived CD20 negative plasma cells. Safety Outcomes The safety profile of cibisatamab plus FAP-4-1BBL was manageable, with most adverse events (AEs) occurring early. Participants routinely received low-dose corticosteroid and non-steroidal anti-inflammatory premedication before infusions. The highest doses administered were 130 mg QW and 90 mg Q3W. Two DLTs were reported: presyncope (Grade 3 SAE, related to FAP-4-1BBL) in the FAP-4-1BBL QW 35 mg plus cibisatamab 100 mg cohort, and cytokine release syndrome (CRS; Grade 3 SAE, related to FAP-4-1BBL and cibisatamab) in the FAP-4-1BBL QW 50 mg plus cibisatamab 100 mg cohort. Overall safety is summarized in Supplementary Table 2. All participants (52/52) experienced at least one AE. The most common AEs (> 30%) were fatigue (33 participants, 63.5%), CRS (30, 57.7%), diarrhea (29, 55.8%), pyrexia (26, 50.0%), decreased appetite (23, 44.2%), cough (17, 32.7%), anemia (17, 32.7%), nausea (16, 30.8%) and arthralgia (16, 30.8%) (Supplementary Table 3). Grade ≥ 3 AEs occurred in 38 participants (73.1%), and SAEs in 34 participants (65.4%). Four participants (7.7%) had Grade 5 AEs (sepsis, sudden death, general physical health deterioration, and cytomegalovirus (CMV)-related colitis); the latter two were considered related to study treatment. CRS occurred predominantly during Cycle 1 (median onset Day 2) and was Grade 1 in 25 participants (48.1%), Grade 2 in 3 (5.8%), and Grade 3 in 2 (3.8%). Serious CRS events occurred in 13 participants (25%) overall. In cohorts treated with cibisatamab 60 mg in Cycle 1 followed by 100 mg from Cycle 2, serious CRS occurred in 4/27 participants (14.8%). CRS events are summarized in Supplementary Table 4. Gastrointestinal toxicity, a known on-target/off-tumor effect of cibisatamab 16 , was also observed with combination therapy: nausea (30.5%), vomiting (23.1%), and diarrhea (55.8%). Diarrhea occurred throughout treatment (median onset, 36 days; range, − 2 to 393 days relative to first dose). The negative onset reflects the study design, in which treatment was initiated on study day 1, while some participants received obinutuzumab lead-in dosing on study day − 13 or − 8, depending on the protocol version. Colitis was reported in 7 participants (13.5%), including colitis, immune-mediated enterocolitis, and CMV colitis. Pharmacokinetics and Immunogenicity Outcomes FAP-4-1BBL exposure increased with dose. Geometric mean PK parameters, including AUC and Cmax, rose with escalating doses. Elimination was non-linear, consistent with target-mediated drug disposition (TMDD) 20 . PK parameters are summarized in Supplementary Table 5. Cibisatamab exposure was consistent with prior reports in monotherapy and in combination with atezolizumab, with Cycle 1 Cmax and AUC within established ranges 16 , suggesting no meaningful impact of FAP-4-1BBL co-administration on cibisatamab PK. Cycle 1 PK parameters are summarized in Supplementary Table 6. Two participants pretreated with obinutuzumab developed ADAs to FAP-4-1BBL; no participants had treatment-induced ADAs to cibisatamab. Efficacy Outcomes Response Rates Fifty-one of 52 participants were evaluable for response (one participant had no on-treatment tumor assessment owing to clinical deterioration). Partial responses (PR) by RECIST 1.1 were observed in 9 participants (17.6%, 95% CI: 6.06–28.55), including 7 confirmed PRs (13.7%, 95% CI: 3.22–23.70). No complete responses were observed. PRs occurred in 7/29 participants (24.1%) receiving FAP-4-1BBL QW and 2/22 (9.1%) receiving FAP-4-1BBL Q3W. By cibisatamab dosing strategy, PRs occurred in 4/25 participants (16.0%) treated with 100 mg from Cycle 1 and 5/26 (19.2%) treated with cibisatamab 60→100 mg (step-up). In the highest targeted QW dose groups (90 mg and 130 mg combined; n = 11; all treated with cibisatamab 60→100 mg), PRs occurred in 4 participants (36.4%), including 3 confirmed PRs (27.3%). In the highest targeted Q3W dose group (90 mg; n = 3; all treated with cibisatamab 60→100 mg), no PRs were observed (limited enrollment due to strategic considerations). The disease control rate (DCR) was 50.9% (26/51), including 58.6% (17/29) with FAP-4-1BBL QW and 40.9% (9/22) with FAP-4-1BBL Q3W. DCR was 56.0% (14/25) in participants treated with cibisatamab 100 mg from Cycle 1 and 46.1% (12/26) in participants treated with cibisatamab 60→100 mg, irrespective of FAP-4-1BBL schedule. In the highest targeted QW dose groups (90 mg and 130 mg combined; n = 11), DCR was 63.6%. Cohort-level efficacy is summarized in Table 1 , categorized by FAP-4-1BBL dose and schedule and by cibisatamab dosing strategy (cibisatamab 100 mg versus cibisatamab 60→100 mg). Progression-free survival (PFS) by investigator assessment suggested a trend toward longer disease control in participants treated in the later dose-escalation cohorts Part 1 Cohort 4 (P1C4) and Part 1 Cohort 5 (P1C5) (n = 12), compared with earlier Part 1 Cohorts 1–3 (P1C1–P1C3) (n = 18) (Supplementary Fig. 3). Part 1 Cohorts 1–3 evaluated earlier FAP-4-1BBL dose levels with cibisatamab administered at a flat 100-mg dose starting from Cycle 1, whereas Part 1 Cohorts 4 and 5 explored higher and optimized FAP-4-1BBL dose levels in combination with a refined cibisatamab step-up regimen (60 mg at Cycle 1 followed by 100 mg from Cycle 2). Although progression events occurred in both groups, separation of the PFS curves became apparent at later timepoints, with a higher proportion of participants in Part 1 Cohorts 4–5 remaining progression-free beyond approximately 150 days. Median PFS was not reached in the Part 1 Cohort 4–5 group at the time of analysis, whereas participants in earlier Part 1 cohorts experienced progression earlier. These findings are exploratory and should be cautiously interpreted given the limited sample size and non-randomized cohort structure. Table 1. Best overall response by investigator (RECIST v1.1) Cibisatamab + FAP-4-1BBL. Confirmed and unconfirmed responses (left) and confirmed responses only (right). Combination schedule FAP-4-1BBL dose (mg) Cibisatamab dose (mg) n evaluable PR n (%) (95% CI) SD n (%) (95% CI) ORR n (%) (95% CI) DCR n (%) Confirmed PR n (%) (95% CI) Confirmed SD n (%) (95% CI) Confirmed ORR n (%) (95% CI) Confirmed DCR n (%) FAP-4-1BBL QW + Cibisatamab Q3W 35 100 12** 2 (16.7) (0.00–41.92) 5 (41.7) (9.61–73.73) 2 (16.7) (0.00–41.92) 7 (58.3) 2 (16.7) (0.00–41.92) 5 (41.7) (9.61–73.73) 2 (16.7) (0.00–41.92) 7 (58.3) 50 100 3 0 (0.00–16.67) 1 (33.3) (0.00–100.00) 0 (0.00–16.67) 1 (33.3) 0 (0.00–16.67) 1 (33.3) (0.00–100.00) 0 (0.00–16.67) 1 (33.3) 50 60→100 3 1 (33.3) (0.00–100.00) 1 (33.3) (0.00–100.00) 1 (33.3) (0.00–100.00) 2 (66.6) 1 (33.3) (0.00–100.00) 1 (33.3) (0.00–100.00) 1 (33.3) (0.00–100.00) 2 (66.6) 90 60→100 6* 3 (50.0) (1.66–98.14) 1 (16.7) (0.00–54.82) 3 (50.0) (1.66–98.14) 4 (66.6) 2 (33.3) (0.00–79.39) 2 (33.3) (0.00–79.39) 2 (33.3) (0.00–79.39) 4 (66.6) 130 100 5 1 (20.0) (0.00–65.06) 2 (40.0) (0.00–92.94) 1 (20.0) (0.00–65.06) 3 (60.0) 1 (20.0) (0.00–65.06) 2 (40.0) (0.00–92.94) 1 (20.0) (0.00–65.06) 3 (60.0) FAP-4-1BBL Q3W + Cibisatamab Q3W 35 100 10 2 (20.0) (0.00–49.79) 4 (40.0) (4.64–75.36) 2 (20.0) (0.00–49.79) 6 (60.0) 1 (10.0) (0.00–33.59) 5 (50.0) (14.01–85.99) 1 (10.0) (0.00–33.59) 6 (60.0) 35 60→100 9 0 (0.00–5.56) 2 (22.2) (0.00–54.94) 0 (0.00–5.56) 2 (22.2) 0 (0.00–5.56) 2 (22.2) (0.00–54.94) 0 (0.00–5.56) 2 (22.2) 90 60→100 3 0 (0.00–16.67) 1 (33.3) (0.00–100.00) 0 (0.00–16.67) 1 (33.3) 0 (0.00–16.67) 1 (33.3) (0.00–100.00) 0 (0.00–16.67) 1 (33.3) All All All 51 9 (17.6) (6.20–29.09) 17 (33.3) (19.42–47.25) 9 (17.6) (6.20–29.09) 26 (50.9) 7 (13.7) (3.30–24.15) 19 (37.3) (23.01–51.50) 7 (13.7) (3.30–24.15) 26 (50.9) Abbreviations: PR, partial response; SD, stable disease; ORR, objective response rate; DCR, disease control rate; QW, once weekly; Q3W, once every 3 weeks. Cibisatamab 60→100 indicates 60 mg at Cycle 1 followed by 100 mg from Cycle 2 onward. * Seven participants were enrolled in this cohort; one withdrew due to an adverse event and had no RECIST assessment. ** One participant withdrew after a single study dose (FAP-4-1BBL + cibisatamab) but was later reported as a confirmed PR without receiving any anticancer therapy, per investigator assessment. Treatment duration suggested longer exposure with QW versus Q3W FAP-4-1BBL (Fig. 1 ). For example, median FAP-4-1BBL treatment duration was 123 days (range 1–435) in the 35 mg QW cohort plus cibisatamab 100 mg (n = 12) versus 64 days (range 1–260) in the corresponding 35 mg Q3W cohort (n = 10). Similarly, in the 90 mg cohorts, median duration was 99 days (range 1–232) for QW (n = 6) versus 43 days (range 22–57) for Q3W (n = 3). In cohorts treated with the same FAP-4-1BBL dose and schedule, longer treatment duration was observed in participants who received cibisatamab 60→100 mg versus 100 mg from Cycle 1 (Supplementary Table 7). Exposure-Safety Analysis We evaluated associations between FAP-4-1BBL exposure and AEs of interest (CRS, infusion-related reactions (IRR), rash, diarrhea, and colitis). For acute events (CRS and IRR), analyses used first-dose Cmax (evaluable: QW n = 29, Q3W n = 22). For delayed events (rash, diarrhea, colitis), analyses used cumulative exposure over the first 63 days (AUC0–63 days) to approximate exposure through approach to steady state. No consistent exposure–AE relationships were observed for CRS, IRR, rash, or colitis across regimens. A possible trend between higher exposure and diarrhea was noted in the QW regimen but not in Q3W. Owing to low event counts, logistic regression analyses were underpowered for definitive inference. Additional details are provided in Supplementary Fig. 4. Analysis of Baseline Markers in Association with Clinical Response To explore predictors of response, we assessed best change in SLD across participants (Fig. 2 ). PRs occurred across a range of baseline biomarker profiles, with clinical benefit not restricted to a single subgroup. Heatmap overlays of CEA and FAP expression, immune phenotype, prior therapy, and baseline tumor CD8⁺ infiltration suggested responses also occurred in participants with adverse features, including liver metastases and excluded/desert immune phenotypes. Baseline profiling identified tumor-intrinsic and clinical features associated with outcome. Tumor CEA by IHC was associated with DCR (Fig. 3 A): participants with low CEA (H-score ≤ 100) had shorter PFS than those with medium (101–200) or high (201–300) scores. This association was not reproduced when CEACAM5 was assessed by RNA-seq (Supplementary Fig. 5). Baseline FAP expression did not correlate with DCR or ORR; tertile stratification suggested a favorable PFS trend in the intermediate-expression group (Fig. 3 B). Waterfall plot showing best percentage change from baseline in SLD by RECIST v1.1. Bars are color-coded by best overall response (BOR): PD (red), SD (yellow), and PR (blue). Heatmaps indicate baseline tumor and clinical features, including CEA and FAP expression, immune phenotype (inflamed, excluded, desert), liver metastases, prior lines of therapy, and dosing regimen. Gradients represent quantitative assessment for CEA and FAP H-score and CD8⁺/Ki67⁺ density. Consensus molecular subtype (CMS) analysis 21 suggested enrichment of responders in CMS2, an epithelial, WNT/MYC-driven transcriptional subtype (Fig. 3 C). Although the overall association between CMS category and clinical outcome did not reach statistical significance (P = 0.069), participants with CMS2 tumors showed numerical improvements in both PFS and OS. No partial responses (PRs) were observed in CMS1 or CMS4 tumors. CMS1 tumors, which are enriched for immune activation, interferon signaling, and MSI features and are frequently associated with a more favorable prognosis in early-stage disease, did not derive objective benefit in this heavily pretreated pMMR/MSS metastatic cohort. CMS4 tumors, characterized by epithelial–mesenchymal transition and a TGF-β–driven immunosuppressive phenotype, were likewise non-responsive 22 . Pretreatment immune phenotyping showed that most partial responses (PRs) occurred in tumors with an excluded immune phenotype, whereas progressive disease (PD) was more frequently observed in immune-desert tumors. 23 . The T-cell excluded phenotype was associated with longer PFS (Fig. 3 E). In contrast, neither baseline tumor 4-1BB⁺ or CD8⁺ density nor circulating cytokines/immune-cell subsets (including IL-8, lymphocyte-to-neutrophil ratio, circulating CD8⁺/Ki67⁺ T cells, IFN-γ, CXCL10, soluble IL-2 receptor, and soluble target proteins) were associated with clinical outcomes (Supplementary Fig. 5). Active liver metastases, an adverse prognostic factor 24 , 25 , did not influence response, PFS, or OS; 4 of 7 of the confirmed PRs (57%) occurred in participants with at least one liver target lesion (Fig. 3 D). Pharmacodynamics Pharmacodynamic effects were assessed in peripheral blood and paired tumor biopsies. Pharmacodynamic changes upon treatment in peripheral blood Peripheral blood cytokines and immune-cell populations were analyzed and compared with published cibisatamab monotherapy data 16 , and further extended by additional pharmacodynamic analyses performed on previously unpublished data from that study. Cibisatamab plus FAP-4-1BBL induced increases in circulating IFNγ, soluble CD25 (sCD25), and IL-6 (Fig. 4 ). IFNγ rose by Day 15 and remained elevated; sCD25 showed a similar sustained pattern, with both markers induced more strongly than with cibisatamab alone. IL-6 increased with kinetics similar to monotherapy. TNFα and IL-10 were also higher at Day 15, consistent with an enhanced immunostimulatory profile. Flow cytometry showed an early, transient circulating lymphocyte reduction (8 hours and C1D2), consistent with margination and/or trafficking, followed by return to baseline. The combination increased proliferating CD8⁺Ki67⁺ T cells and activated CD8⁺ subsets (Tim3, 4-1BB, HLA-DR), with expansion of central and effector memory compartments. Compared with cibisatamab monotherapy, Day 22 signs of activation were greater, including higher CD8⁺4-1BB⁺ and CD8⁺HLA-DR⁺ frequencies and increased PD-1 (CD279) expression on central memory and effector CD8⁺ T cells, consistent with more pronounced CD8⁺ activation. Soluble 4-1BB (sCD137) increased across response groups, consistent with target engagement 26 . Participants with PR showed more sustained induction during cycles 1–2, whereas SD/PD showed lower and more variable changes. Serum CEA (sCEA) declined in participants with PR from Cycle 2 onward (median ~ 0.5–1.0 log reductions by cycles 6–8), with smaller changes in SD and little change in PD. Pharmacodynamic changes upon treatment in fresh paired tumor biopsies Paired fresh biopsies were collected at baseline and Day 23 after the first combination dose to assess intratumoral immune pharmacodynamics. Duplex IHC (CD8, Ki67) showed > 2-fold increases in intratumoral CD8⁺ T cells in 12/14 evaluable pairs and > 2-fold increases in CD8⁺Ki67⁺ cells in 10/14 pairs (Fig. 5 A; Supplementary Fig. 6). Immune phenotype profiling 23 classified baseline lesions as desert (79%), excluded (14%) or inflamed (7%). On treatment, desert lesions decreased to 21%, with on-treatment assessed increases in excluded (50%) and inflamed (29%) phenotypes (Fig. 5 B). Quantitative image analysis showed higher post-treatment CD8⁺ and CD8⁺Ki67⁺ densities after combination therapy than after cibisatamab monotherapy: median intratumoral CD8⁺ cell density was approximately threefold higher than that observed with monotherapy (p = 0.056), with a corresponding fivefold increase in CD8⁺Ki67⁺ cell density (p = 0.019) (Fig. 5 c). Such increases were also observed in 10/11 participants with liver metastases (Supplementary Fig. 6). RNA-seq of paired biopsies passing QC (n = 6) showed induction of immune effector genes (CXCL9, CXCL10, IFNγ, TNFRSF9/4-1BB) (Fig. 5 D; Supplementary Fig. 6). IFN-γ response, cytotoxic T-cell, T-effector and antigen processing machinery (APM) signatures increased (all p < 0.01). Induction was greater in PR/SD than PD despite higher baseline expression in PD, consistent with a less robust transcriptional response in PD tumors. A modest but significant increase in the M1/M2 macrophage ratio was estimated based on gene signatures (Supplementary Fig. 6). Adaptive immune-regulatory programs were also induced, including PD-L1/CD274, LAG3, EOMES 27 , 28 and IDO1 29,30 (Supplementary Fig. 6). FAP and CEA expression showed no consistent treatment-induced change (Supplementary Fig. 6). Circulating ctDNA In an exploratory tissue-informed patient-specific multiplex PCR assay, baseline ctDNA levels varied widely (median 1,604 MTM/ml) and were generally lower in participants with PR (with one exception) than in participants with SD or PD (Fig. 6 A). Lower baseline ctDNA showed a non-significant trend toward improved PFS (p = 0.061) and OS (p = 0.2) (Fig. 6 B). By Cycle 3 Day 1 (C3D1), lower residual ctDNA ( 3-fold in 5/5 participants with PR and 6/8 with SD, whereas participants with PD showed minimal change or increases (Fig. 6 C). Greater ctDNA clearance was associated with longer PFS (p = 0.009) (Fig. 6 D). Discussion Our study provides clinical evidence in solid tumors that coordinated delivery of synthetic signal 1 and signal 2 31,32 can be achieved through the combination of a tumor-directed T-cell engager and a spatially restricted costimulatory agonist. As a signal‑1 provider, the CEA-directed T-cell engager cibisatamab induces antigen-dependent T-cell activation and tumor infiltration 14 and has previously demonstrated modest but reproducible clinical activity in heavily pretreated MSS colorectal cancer patients at doses comparable to those evaluated here 16 . Optimal T-cell activation requires integration of TCR–CD3 signaling with costimulatory inputs, including those mediated by 4‑1BB and other members of the TNFR family 33 . Importantly, 4-1BB is selectively upregulated following initial T-cell priming and is undetectable on resting T lymphocytes 11 , 18 . Based on this biology, we hypothesized that coupling cibisatamab-mediated T-cell engagement with FAP‑4‑1BBL–mediated costimulation would amplify antitumor immunity while restricting signal‑2 delivery to the tumor microenvironment, where FAP is selectively expressed by cancer-associated fibroblasts outside contexts of acute tissue repair or scar formation 34 , 35 . In the BP42675 study, the combination of cibisatamab and FAP‑4‑1BBL following obinutuzumab pre-treatment demonstrated a manageable safety profile, robust pharmacodynamic immune activation, and encouraging preliminary efficacy in patients with heavily pretreated MSS mCRC. Obinutuzumab was incorporated to mitigate anti-drug antibody formation previously observed with cibisatamab monotherapy, thereby enabling sustained drug exposure and preservation of biological activity 16 . Although B-cell depletion raises theoretical concerns regarding disruption of tertiary lymphoid structures and humoral support of antitumor immunity 36 – 38 , no clear detriment to clinical or immunologic outcomes was observed in this study. Nevertheless, the absence of a non–B-cell-depleted comparator cohort represents an important limitation and warrants further evaluation since pro-tumor activities of B lymphocytes have also been reported 39 . Dose and schedule selection for cibisatamab was informed by prior clinical experience and by preclinical imaging studies demonstrating prolonged intratumoral retention despite a short systemic half-life 15 . For FAP‑4‑1BBL, both weekly and every‑3‑week dosing regimens were explored to evaluate the impact of sustained versus intermittent costimulatory signaling, an area of ongoing debate for agonist 4‑1BB biology 40 . Notably, numerically higher response rates, deeper tumor shrinkage, and longer treatment durations were observed in cohorts receiving weekly FAP‑4‑1BBL, particularly at higher dose levels, suggesting that more continuous costimulatory engagement may be advantageous in this setting. Across the response-evaluable population, cibisatamab plus FAP‑4‑1BBL achieved numerically higher ORR and DCR than previously reported for cibisatamab monotherapy 16 , with the greatest activity observed in higher-dose FAP‑4‑1BBL weekly cohorts. These findings are consistent with preclinical evidence demonstrating synergistic antitumor effects of T-cell engagement combined with 4‑1BB costimulation in murine models and patient-derived tumor systems 18 , 41 . While cross-trial comparisons should be interpreted cautiously, the convergence of clinical and pharmacodynamic signals supports a true biological interaction between signal‑1 and signal‑2 to achieve a more pronounced T-cell activation and tumor infiltration. The safety profile of the combination was largely consistent with the known toxicities of cibisatamab and with prior experience from the FAP‑4‑1BBL first-in-human study. Gastrointestinal toxicity, a known on-target/off-tumor effect of CEA-directed T-cell engagement 16 , was common but generally manageable and not dose-limiting. Importantly, the addition of tumor-targeted 4‑1BB costimulation did not result in excess systemic immune toxicity, contrasting with historical experience using untargeted 4‑1BB agonist antibodies 42 . Extensive biomarker analyses provided mechanistic context for treatment activity. Analyses were primarily contextualized against published and additional cibisatamab monotherapy data in patients with mCRC 16 , as this represents the most relevant disease-matched comparator. In contrast, the first-in-human FAP-4-1BBL monotherapy study enrolled multiple tumor types, limiting its utility for direct comparison in MSS mCRC 43 . Nevertheless, integration of data across all three studies indicates that the cibisatamab plus FAP-4-1BBL combination induced the largest increases in intratumoral CD8⁺ and proliferating CD8⁺Ki67⁺ T cells, accompanied by concordant peripheral immune activation 44 . In BP42675, these effects were observed even in liver metastases, suggesting partial mitigation of liver-associated immune suppression. Baseline biomarker analyses identified tumor CEA protein expression by immunohistochemistry as a potential predictor of disease control and progression-free survival. Intermediate FAP expression was associated with more favorable outcomes, suggesting that effective costimulation may require sufficient—but not excessive—CAF targeting 45 . Responses were enriched in CMS2 tumors and in lesions with an immune-excluded phenotype, indicating that tumors traditionally considered poorly inflamed may still be amenable to T-cell engager–based strategies when combined with tumor-localized artificial costimulation 46 , 47 . On-treatment biomarkers further underscored the depth of immune activation achieved with the combination, including sustained induction of IFNγ, soluble CD25, activated CD8⁺ T-cell subsets, and soluble CD137, consistent with engagement of the 4-1BB pathway 26 . A transient early decrease in circulating lymphocyte counts was observed within hours of dosing, consistent with lymphocyte margination and rapid immune-cell redistribution, a phenomenon previously described for T-cell engager therapies 7 , 16 , 48 , 49 . In addition, early reductions in serum CEA and circulating tumor DNA (ctDNA) were associated with improved clinical outcomes, supporting their utility as minimally invasive pharmacodynamic and early response biomarkers 50 – 52 . In the context of robust and sustained immune activation, treatment was accompanied by increased expression of several inhibitory checkpoint molecules and metabolic regulators, including PD-L1, LAG3, EOMES, and IDO1, consistent with adaptive feedback mechanisms that can emerge under heightened immune stimulation 53 . Notably, these transcriptional changes occurred alongside a treatment-associated shift toward a more pro-inflammatory myeloid milieu, reflected, for example, by an increased M1 to M2 macrophage polarization signature ratio and upregulation of antigen presentation machinery signatures, which was not observed with cibisatamab monotherapy 16 , and accompanied by marked CD8⁺ T-cell infiltration and activation. Together, these findings indicate that immune activation predominates over counter-regulatory signals and support the biological activity of the combination, while suggesting that simultaneous or sequential co-treatment with PD-(L)1 checkpoint inhibitors may represent a rational strategy to enhance response durability 43 . In conclusion, this study establishes a clinically tractable framework for coordinated delivery of signal-1 and signal-2 immunostimulation within the tumor microenvironment of solid tumors. By pairing a tumor-directed T-cell engager with spatially restricted 4-1BB costimulation, cibisatamab plus FAP-4-1BBL achieved enhanced intratumoral T-cell activation, encouraging antitumor activity, and a manageable safety profile in a disease historically refractory to immunotherapy. These data provide proof of principle that localized costimulation can amplify the efficacy of T-cell engagers without recapitulating the systemic toxicities associated with untargeted agonists. Beyond MSS colorectal cancer, this strategy offers a modular and broadly applicable paradigm for overcoming immune exclusion and resistance in non-inflamed tumors and supports further clinical development of T-cell engager–based combinations incorporating tumor-restricted costimulatory signals to improve the depth and durability of response. Methods Inclusion and ethics This study was an open-label, multicenter, phase 1b dose-escalation trial conducted across 15 globally approved clinical sites, of which 12 were actively recruiting participants. The trial was performed in accordance with Good Clinical Practice guidelines and applicable regulatory requirements. Approval was obtained from institutional review boards or independent ethics committees at each participating site. All participants provided written informed consent prior to enrolment and before the initiation of any study-related procedures. Trial design, participants and treatments The trial consisted of two sequential dose-escalation parts designed to explore different FAP-4-1BBL exposure profiles. Part 1 evaluated continuous exposure to FAP-4-1BBL administered weekly (QW), whereas Part 2 investigated a pulsatile exposure strategy with FAP-4-1BBL administered every 3 weeks (Q3W In both study parts, cibisatamab was administered intravenously Q3W, with dosing consisting of either 100 mg in all cycles or a step-up regimen of 60 mg in cycle 1 followed by 100 mg thereafter. (Supplementary Figures 1 and 2). At study initiation, participant eligibility required high tumor CEACAM5 expression determined by quantitative PCR analysis of FFPE tumor tissue, consistent with prior cibisatamab studies 16 . Following enrollment of the first 11 participants, this requirement was removed to enable assessment of the cibisatamab and FAP-4-1BBL combination in an unselected colorectal cancer population, reflecting the high prevalence of CEA expression in CRC 16 . To mitigate the development of anti-drug antibodies (ADAs) against cibisatamab, as previously reported 16 , all participants received B-cell–depleting obinutuzumab as pre-treatment 1–2 weeks prior to Cycle 1 Day 1. Obinutuzumab dosing was repeated after 6 months of treatment. Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 5.0. Cytokine release syndrome (CRS) was graded according to the American Society for Transplantation and Cellular Therapy consensus criteria. Study objectives and corresponding endpoints for BP42675 are summarized in Supplementary Table 8. Statistical analysis All 52 enrolled participants were evaluable for safety and dose-limiting toxicities (DLTs). One participant did not undergo an on-treatment tumor assessment and was therefore excluded from efficacy analyses, resulting in an efficacy-evaluable population of 51 participants. All analyses were exploratory in nature. PK sampling and parameter calculation Serial pharmacokinetic (PK) samples were collected from all participants to characterize the PK profiles of FAP-4-1BBL and cibisatamab following initial administration. Based on these data, key PK exposure parameters were calculated for each participant using standard non-compartmental analysis (NCA). Only participants with a sufficient number of PK samples to allow robust NCA were included in exposure–response analyses. Exposure–safety analysis Potential associations between FAP-4-1BBL exposure and the incidence and severity of adverse events of particular interest were evaluated in all safety-evaluable participants. Adverse events assessed included CRS, infusion-related reactions (IRR), rash, diarrhea, and colitis. The PK exposure metric selected for analysis was tailored to the anticipated timing and mechanism of each adverse event. For acute events such as CRS and IRR, which typically occur shortly after infusion, exposure was evaluated using maximum plasma concentration (C_max) following the first dose. For adverse events with potentially delayed onset, including rash, diarrhea, and colitis, cumulative exposure over the first 63 days (AUC_0–63 days) was used, corresponding to the period during which drug concentrations were expected to approach steady state. Exposure–safety analyses were stratified by QW and Q3W dosing regimens. Exploratory box plots were generated to visualize the distribution of exposure metrics across adverse event grades. Baseline and pharmacodynamic assessments Pharmacodynamic assessments were conducted using whole blood, plasma, and tumor tissue samples, all of which were analyzed centrally. Whole blood and plasma samples were collected to evaluate changes in soluble biomarkers, cytokines, and immune-cell subsets using validated flow cytometry and immunoassay methods. Tumor tissue samples were analyzed by immunohistochemistry (IHC), immunofluorescence, and whole-transcriptome RNA sequencing to assess biomarkers of tumor inflammation and immune infiltration, including CD8⁺ tumor-infiltrating lymphocytes (TILs) and immune-related gene signatures. On-treatment tumor biopsies for the combination study were obtained on Cycle 2 Day 2 (Day 23). In the cibisatamab monotherapy study, on-treatment biopsies were collected at either Cycle 2 Day 1 (Day 21; Q3W dosing) or Cycle 7 Day 1 (Day 43; QW dosing), reflecting differences in treatment schedules. Cytokines, soluble 4-1BB and CEA assessment Blood samples were collected at baseline and at 1, 8, 15, 22, 29, 36, and 43 days following the first administration of FAP-4-1BBL. Peripheral inflammatory cytokines, soluble IL-2 receptor (sCD25), soluble 4-1BB, and serum CEA were quantified using validated ELISA-based platforms, including ProteinSimple ELLA cartridges for IFN-γ, IL-6, IL-2 receptor, and TNF-α, Peprotech ELISA for soluble 4-1BB, and the Elecsys CEA assay (Cobas E170). For cytokine analyses, values below the lower limit of quantification (LLOQ) were imputed as 0.5 × LLOQ with the addition of a small random noise term to facilitate model convergence, followed by log2 transformation (log2[value + 1]). Soluble CEA values were analyzed without imputation and were log10-transformed. Soluble 4-1BB values below LLOQ were handled similarly to cytokines and log10-transformed. All analyses were performed on changes from baseline. Random-effects models were used for longitudinal analyses, with participant specified as a random effect and visit timepoint as a fixed effect. Correction for multiple testing was performed using the false discovery rate (FDR), with statistical significance defined as FDR < 0.05. Flow cytometry Peripheral blood samples were collected in sodium heparin tubes and processed according to validated assay protocols. All staining and incubation steps were performed at room temperature in the dark. Data acquisition was performed using BD FACSCanto II instruments, with analysis conducted using FACSDiva software and assay-specific acquisition templates. Cell preparation was automated using a BD FACS Lyse Wash Assistant. Surface staining was performed using antibody cocktails applied to whole blood samples, followed by red blood cell lysis, washing, and resuspension in phosphate-buffered saline. For intracellular staining, samples were additionally permeabilized using Perm Buffer II prior to incubation with intracellular antibody cocktails. A complete list of antibody reagents is provided in Supplementary Table 9. Immunohistochemistry Tumor tissue was collected at baseline and on Cycle 2 Day 2 (±2 days where feasible), preferentially from the same lesion. All samples were processed and paraffin-embedded according to standardized histopathology protocols. FFPE tissue sections were stained for hematoxylin and eosin, IHC, and immunofluorescence at Discovery Life Sciences (Kassel, Germany) or Roche Tissue Diagnostics (Tucson, Arizona). Only samples meeting predefined quality-control criteria for tumor content and tissue integrity were included. Details of staining assays are provided in Supplementary Table 10. CD3/CD8/Ki67/4-1BB/OX40 multiplex immunofluorescence Multiplex immunofluorescence staining using five primary antibodies and tyramide signal amplification fluorophores was performed on a Ventana DISCOVERY ULTRA IHC/ISH platform, as previously described 54 . Whole-slide imaging was performed using a Zeiss Axioscan Z1 scanner. Quantitative image analysis was conducted using a custom-developed algorithm implemented in HALO (Indica Labs). Antibody and fluorophore details are provided in Supplementary Table 11. Statistics for immunohistochemistry and immunofluorescence analyses Pharmacodynamic analyses were restricted to participants with matched baseline and on-treatment tumor samples. Paired changes were assessed using Wilcoxon matched-pairs signed-rank tests. Comparisons between pooled CR/PR participants and SD/PD participants were conducted using unpaired Mann–Whitney tests. Additional comparisons were performed between participants with progressive disease and all remaining participants. All analyses were performed using GraphPad Prism version 10.0.0. Statistics for baseline associations with response Associations between baseline biomarkers and clinical response were evaluated using unpaired Mann–Whitney tests for continuous variables. Fisher’s exact test was used to assess associations between liver metastases, consensus molecular subtype (CMS), and objective response. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan–Meier methods with 95% confidence intervals, and group comparisons were performed using log-rank tests. Gene expression analyses RNA sequencing and processing RNA was isolated from FFPE tumor tissue using the Qiagen AllPrep DNA/RNA FFPE kit. Library preparation was performed using the Illumina TruSeq RNA Exome kit. Base calling was conducted using bcl2fastq2, and sequence quality was assessed using FastQC 55 . Paired-end reads were aligned to the human genome (hg38) using STAR. Quality control was performed using MultiQC 54 . Gene-level read counts were generated using featureCounts 56 and normalized to transcripts per million (TPM). Differential gene expression and signature analyses Differential gene expression analyses were conducted using the limma package with log-transformed TPM values. Single-sample gene signature scores were calculated as the mean Z score of constituent genes. For signatures with up- and downregulated components, Z scores for downregulated genes were sign-inverted. Group comparisons were performed using limma. Adjustment of p-values Raw p-values were adjusted for multiple testing using the Benjamini–Hochberg procedure 57 . ctDNA analysis Circulating tumor DNA was analyzed using a personalized, tumor-informed, 16-plex PCR next-generation sequencing assay (Signatera RUO, Natera) 58 . Patient-specific somatic variants were identified from whole-exome sequencing data and used to design multiplex PCR assays. Plasma samples with ≥2 detected variants were classified as ctDNA-positive, and ctDNA levels were reported as mean tumor molecules per milliliter (MTM/ml). Whole-exome sequencing Whole-exome sequencing was performed on DNA extracted from macrodissected FFPE tumor tissue and matched germline DNA from peripheral blood, as previously described 58 . Reads were aligned to hg38 using BWA, duplicates were removed, and base quality score recalibration was performed using GATK 59 . Somatic variants were identified using Mutect2 and annotated with VEP 60 . Tumor mutational burden was calculated after standard filtering and normalized per megabase 61,62 . ctDNA data analysis and statistics Kaplan–Meier analyses for PFS and OS were performed using survfit2 and visualized using ggsurvfit. Participants were stratified into high and low ctDNA groups based on median baseline values. Landmark analyses at Cycle 3 Day 1 were applied to minimize immortal-time bias, with follow-up re-anchored at the landmark timepoint. Declarations Acknowledgements and Disclosures Acknowledgements We thank the patients for their participation in the study; the study nurses and site coordinators for their professionalism and dedication in supporting the conduct of the study; Christina Claus, Christian Klein, Vaios Karanikas (Roche Innovation Center Zurich) and Marie-Hélène Wasmer (Roche Innovation Center Basel) for continuous support and valuable discussions; Alfred Geiger (pRED Basel) for ensuring drug availability; Marleen Wilde, Nadine Kumpesa, and Marine Cordonnier (pRED Basel) for their contributions to study set-up; and Chiahuey Oii (pRED Basel) for providing the initial RNA-seq analysis. Sietzke Heyn for editorial assistance with the preparation of this article; and finally the numerous anonymous persons at the research sites as well as at Roche that have also been essential in supporting this project. Medical writing and editorial assistance were provided by Dr. Raja Choudhury in accordance with Good Publication Practice (GPP3) guidelines and was funded by F. Hoffmann-La Roche Ltd. Author contributions I.M, E.C.A, C.Q, M.D, I.B, S.B, F.T, M.C.R, M.M.G, V.W., T.W.K, V.M, were clinical investigators on the study. A.B, E.G, C.H, T.T, H.H, C.M, P.U and I.M planned, designed and refined the study. A.B. lead the Roche study team. T.T., E.G and A.B wrote the initial draft of the paper, and all authors reviewed and edited the paper. C.H, I.D and T.T conducted biomarker analyses. C.M conducted pharmacokinetics analysis. L.C. developed the clinical analysis plan and analyzed data. H.H conducted the safety analysis. E.G., E.C., and O.K. provided medical support and oversight. T.S. had the operations oversight Funding This work was supported by F. Hoffmann-La Roche Ltd, which significantly contributed to the study design in collaboration with study investigators, contributed to the collection of data, significantly contributed to the analysis and interpretation of the data in collaboration with study investigators, contributed to the writing of the report with study collaborators, and agreed with the study investigators in the decision to submit the paper for publication. No grant number is applicable. Competing interests Several authors are employees of F. Hoffmann-La Roche Ltd and/or Genentech, Inc., and may hold Roche stock or stock options as part of their employment compensation. Iosune Baraibar reports accommodation and travel expenses from Amgen, Merck, Sanofi, and Servier; and personal speaker honoraria from Astra Zeneca and Amgen. Victor Morena declared financial and non-financial interests with Abbvie, Achilles, Adaptimmune, Adc Therapeutics, Ascendis Pharma, Astrazeneca, Bayer, Beigene, Bicycle Tx, Bioinvent, Biomea Fusion, Biontech, Bms,, Boehringer, C4 Therapeutics, Calico Life Sciences Llc, Celgene, Constellation, Crescendo Biologics, Cullinan, Daiichi Sankyo, Debiopharm, Dragonfly, Enliven Therapeutics, Epizyme, Exelixis, Famewave, F-Star Beta Limited, Genentech, Genmab, Gilead, Grey Wolf Therapeutics, Gsk, Hexal Ag & Sandoz, Hifibio, Hookipa Biotech, Hutchmed, Igm Biosciences, Imcheck Therapeutics, Immunocore, Immutep, Incyte Iomx Therapeutics, Iovance, Italfarmaco, Iteos, Janssen, Light Chain Bioscience, Lilly, Loxo Oncology, Merck, Merus, Miltenyi Biomedicine, Monta Biosciences, Msd, Mythic Therapeutics, Ningbo Newbay, Novartis, Oxford Biotherapeutics, Pfizer, Pharmamar, Pmv Pharma, Prelude Therapeutics Inc, Pyxis Oncology, Regeneron, Relay Terapeutics, Repare Therapeutics, Revolution, Roche, Schrödinger, Scorpion Therapeutics, Seagen, Shattucks, Synthorx, Takeda, Tango Therapeutics, Tesaro, Totus Medicines, Turning Point Therapeutics, Vividion Therapeutics. Fiona Thistlethwaite declared institutional funding from Achilles Ltd, Adaptimmune, Amgen, Biontech, BMS, Byondis, Chugai, Corbus, Crescendo Biologics, Grey Wolf Therapeutics, GSK, Immunocore, Iovance, Janssen, Kymab Ltd/Sanofi, Leucid, Moderna, Novalgen, Nucana, Oxford Vacmedix, Roche, RS Oncology LLC, Seagen, Takeda, T-Knife Therapeutics, UCB, Zelluna, Zymeworks. Advisory board/consultancy/honoraria: AstraZeneca, Grey Wolf Therapeutics, OncoBayes, T-Knife Therapeutics, Waypoint DSMB: Immatics Ignacio Melero declared the following interests: Grants: Roche, AstraZenca, Bristol Myers, Genmab. Consultancy: Roche, Genmab, Regeneron, Pioneer medicines, Bright peak, Pharmamar, Curon, Mestag, Light chain, Emiliano Calvo reports the following interests: ownership interests in Oncoart Associated; honoraria from HM Hospitales Group; had consulting/advisory role at Nanobiotix, Janssen-Cilag, Roche/Genentech, TargImmune Therapeutics, Servier, Bristol-Myers Squibb, Amunix, Adcendo, Anaveon, AstraZeneca/MedImmune, Chugai Pharma, MonTa, MSD Oncology, Nouscom, Novartis, OncoDNA, T-Knife, Elevation Oncology, PharmaMar, Ellipses Pharma, Syneos Health, Genmab and Diaccurate; is president and founder of Foundation INTHEOS. Tae Won Kim declared institutional funding from Genentech and Inocras. Seung Hoon Beom declared the following interests: Researcher/Grant/Contract (clinical trial PI): Roche, MSD, Bayer, Boryung, Takeda, Jeil Pharmaceutical, Merck, Janssen, Pfizer, Amgen, Revolution Medicines, IGM Biosciences, STcube References Alouani, E., et al. Efficacy of immunotherapy in mismatch repair-deficient advanced colorectal cancer in routine clinical practice. An AGEO study. ESMO Open 8, 101574 (2023). Chen, E. & Zhou, W. Immunotherapy in microsatellite-stable colorectal cancer: Strategies to overcome resistance. Crit Rev Oncol Hematol 212, 104775 (2025). Cai, L., Chen, A. & Tang, D. A new strategy for immunotherapy of microsatellite-stable (MSS)-type advanced colorectal cancer: Multi-pathway combination therapy with PD-1/PD-L1 inhibitors. Immunology 173, 209–226 (2024). Morano, F., et al. Temozolomide Followed by Combination With Low-Dose Ipilimumab and Nivolumab in Patients With Microsatellite-Stable, O(6)-Methylguanine-DNA Methyltransferase-Silenced Metastatic Colorectal Cancer: The MAYA Trial. J Clin Oncol 40, 1562–1573 (2022). Hutchings, M., et al. Glofitamab, a Novel, Bivalent CD20-Targeting T-Cell-Engaging Bispecific Antibody, Induces Durable Complete Remissions in Relapsed or Refractory B-Cell Lymphoma: A Phase I Trial. J Clin Oncol 39, 1959–1970 (2021). Abramson, J.S., et al. Glofitamab plus gemcitabine and oxaliplatin (GemOx) versus rituximab-GemOx for relapsed or refractory diffuse large B-cell lymphoma (STARGLO): a global phase 3, randomised, open-label trial. Lancet 404, 1940–1954 (2024). Paz-Ares, L., et al. Tarlatamab, a First-in-Class DLL3-Targeted Bispecific T-Cell Engager, in Recurrent Small-Cell Lung Cancer: An Open-Label, Phase I Study. J Clin Oncol 41, 2893–2903 (2023). Mountzios, G., et al. Tarlatamab in Small-Cell Lung Cancer after Platinum-Based Chemotherapy. N Engl J Med 393, 349–361 (2025). Chen, L. & Flies, D.B. Molecular mechanisms of T cell co-stimulation and co-inhibition. Nat Rev Immunol 13, 227–242 (2013). Chester, C., Sanmamed, M.F., Wang, J. & Melero, I. Immunotherapy targeting 4-1BB: mechanistic rationale, clinical results, and future strategies. Blood 131, 49–57 (2018). Melero, I., et al. CD137 (4-1BB)-Based Cancer Immunotherapy on Its 25th Anniversary. Cancer Discov 13, 552–569 (2023). Melero, I., et al. Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors. Nat Med 3, 682–685 (1997). Tiernan, J.P., et al. Carcinoembryonic antigen is the preferred biomarker for in vivo colorectal cancer targeting. Br J Cancer 108, 662–667 (2013). Bacac, M., et al. A Novel Carcinoembryonic Antigen T-Cell Bispecific Antibody (CEA TCB) for the Treatment of Solid Tumors. Clin Cancer Res 22, 3286–3297 (2016). Lehmann, S., et al. In Vivo Fluorescence Imaging of the Activity of CEA TCB, a Novel T-Cell Bispecific Antibody, Reveals Highly Specific Tumor Targeting and Fast Induction of T-Cell-Mediated Tumor Killing. Clin Cancer Res 22, 4417–4427 (2016). Segal, N.H., et al. CEA-CD3 bispecific antibody cibisatamab with or without atezolizumab in patients with CEA-positive solid tumours: results of two multi-institutional Phase 1 trials. Nat Commun 15, 4091 (2024). Teijeira, A., et al. Three-dimensional colon cancer organoids model the response to CEA-CD3 T-cell engagers. Theranostics 12, 1373–1387 (2022). Claus, C., et al. Tumor-targeted 4-1BB agonists for combination with T cell bispecific antibodies as off-the-shelf therapy. Sci Transl Med 11(2019). Peters, S., et al. Obinutuzumab Pretreatment as a Novel Approach to Mitigate Formation of Anti-Drug Antibodies Against Cergutuzumab Amunaleukin in Patients with Solid Tumors. Clin Cancer Res 30, 1630–1641 (2024). Sanchez, J., et al. Combining mathematical modeling, in vitro data and clinical target expression to support bispecific antibody binding affinity selection: a case example with FAP-4-1BBL. Front Pharmacol 15, 1472662 (2024). Guinney, J., et al. The consensus molecular subtypes of colorectal cancer. Nat Med 21, 1350–1356 (2015). Ciardiello, F., et al. Clinical management of metastatic colorectal cancer in the era of precision medicine. CA Cancer J Clin 72, 372–401 (2022). Hegde, P.S., Karanikas, V. & Evers, S. The Where, the When, and the How of Immune Monitoring for Cancer Immunotherapies in the Era of Checkpoint Inhibition. Clin Cancer Res 22, 1865–1874 (2016). Misiakos, E.P., Karidis, N.P. & Kouraklis, G. Current treatment for colorectal liver metastases. World J Gastroenterol 17, 4067–4075 (2011). Yu, J., et al. Liver metastasis restrains immunotherapy efficacy via macrophage-mediated T cell elimination. Nat Med 27, 152–164 (2021). Glez-Vaz, J., et al. Soluble CD137 as a dynamic biomarker to monitor agonist CD137 immunotherapies. J Immunother Cancer 10(2022). Georganaki, M., et al. Tumor endothelial cell up-regulation of IDO1 is an immunosuppressive feed-back mechanism that reduces the response to CD40-stimulating immunotherapy. Oncoimmunology 9, 1730538 (2020). Jenkins, E., Whitehead, T., Fellermeyer, M., Davis, S.J. & Sharma, S. The current state and future of T-cell exhaustion research. Oxf Open Immunol 4, iqad006 (2023). Hornyák, L., et al. The Role of Indoleamine-2,3-Dioxygenase in Cancer Development, Diagnostics, and Therapy. Front Immunol 9, 151 (2018). Munn, D.H. & Mellor, A.L. Indoleamine 2,3-dioxygenase and tumor-induced tolerance. J Clin Invest 117, 1147–1154 (2007). Bretscher, P.A. The history of the two-signal model of lymphocyte activation: A personal perspective. Scand J Immunol 89, e12762 (2019). Etxeberria, I., et al. Engineering bionic T cells: signal 1, signal 2, signal 3, reprogramming and the removal of inhibitory mechanisms. Cell Mol Immunol 17, 576–586 (2020). Watts, T.H., Yeung, K.K.M., Yu, T., Lee, S. & Eshraghisamani, R. TNF/TNFR Superfamily Members in Costimulation of T Cell Responses-Revisited. Annu Rev Immunol 43, 113–142 (2025). Dziadek, S., et al. Comprehensive analysis of fibroblast activation protein expression across 23 tumor indications: insights for biomarker development in cancer immunotherapies. Front Immunol 15, 1352615 (2024). Kraxner, A., et al. Investigating the complex interplay between fibroblast activation protein α-positive cancer associated fibroblasts and the tumor microenvironment in the context of cancer immunotherapy. Front Immunol 15, 1352632 (2024). Cabrita, R., et al. Tertiary lymphoid structures improve immunotherapy and survival in melanoma. Nature 577, 561–565 (2020). Conejo-Garcia, J.R., Biswas, S., Chaurio, R. & Rodriguez, P.C. Neglected no more: B cell-mediated anti-tumor immunity. Semin Immunol 65, 101707 (2023). Fridman, W.H., et al. Tertiary lymphoid structures and B cells: An intratumoral immunity cycle. Immunity 56, 2254–2269 (2023). de Visser, K.E., Korets, L.V. & Coussens, L.M. De novo carcinogenesis promoted by chronic inflammation is B lymphocyte dependent. Cancer Cell 7, 411–423 (2005). Menk, A.V., et al. 4-1BB costimulation induces T cell mitochondrial function and biogenesis enabling cancer immunotherapeutic responses. J Exp Med 215, 1091–1100 (2018). Trüb, M., et al. Fibroblast activation protein-targeted-4-1BB ligand agonist amplifies effector functions of intratumoral T cells in human cancer. J Immunother Cancer 8(2020). Segal, N.H., et al. Results from an Integrated Safety Analysis of Urelumab, an Agonist Anti-CD137 Monoclonal Antibody. Clin Cancer Res 23, 1929–1936 (2017). Melero, I., et al. A first-in-human study of the fibroblast activation protein-targeted, 4-1BB agonist RO7122290 in patients with advanced solid tumors. Sci Transl Med 15, eabp9229 (2023). Melero, I., et al. 151P Enhanced pharmacodynamic effects upon combination of cibisatamab and FAP-4-1BBL in 3L+ mMSS CRC patients. Annals of Oncology 35, S275-S276 (2024). Lee, H.H. & Al-Ogaili, Z. Fibroblast activation protein and the tumour microenvironment: challenges and therapeutic opportunities. Oncol Rev 19, 1617487 (2025). Melero, I., et al. Amplification of tumor immunity by gene transfer of the co-stimulatory 4-1BB ligand: synergy with the CD28 co-stimulatory pathway. Eur J Immunol 28, 1116–1121 (1998). Ye, Z., et al. Gene therapy for cancer using single-chain Fv fragments specific for 4-1BB. Nat Med 8, 343–348 (2002). Bröske, A.E., et al. Pharmacodynamics and molecular correlates of response to glofitamab in relapsed/refractory non-Hodgkin lymphoma. Blood Adv 6, 1025–1037 (2022). O'Brien, N., et al. T cell margination: investigating the detour of T cells following forimtamig treatment in humanized mice. MAbs 17, 2440578 (2025). Strimpakos, A.S., et al. The impact of carcinoembryonic antigen flare in patients with advanced colorectal cancer receiving first-line chemotherapy. Ann Oncol 21, 1013–1019 (2010). Callesen, L.B., et al. Circulating tumour DNA and its clinical utility in predicting treatment response or survival in patients with metastatic colorectal cancer: a systematic review and meta-analysis. Br J Cancer 127, 500–513 (2022). Tie, J., et al. Circulating Tumor DNA Analyses as Markers of Recurrence Risk and Benefit of Adjuvant Therapy for Stage III Colon Cancer. JAMA Oncol 5, 1710–1717 (2019). Sharma, P., Hu-Lieskovan, S., Wargo, J.A. & Ribas, A. Primary, Adaptive, and Acquired Resistance to Cancer Immunotherapy. Cell 168, 707–723 (2017). Ewels, P., Magnusson, M., Lundin, S. & Käller, M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics 32, 3047–3048 (2016). Andrews, S., Gilley, J. & Coleman, M.P. Difference Tracker: ImageJ plugins for fully automated analysis of multiple axonal transport parameters. J Neurosci Methods 193, 281–287 (2010). Liao, Y., Smyth, G.K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930 (2014). Benjamini, Y. & Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological) 57, 289–300 (2018). Reinert, T., et al. Analysis of Plasma Cell-Free DNA by Ultradeep Sequencing in Patients With Stages I to III Colorectal Cancer. JAMA Oncol 5, 1124–1131 (2019). Auwera, G.v.d. & O'Connor, B.D. Genomics in the cloud: using Docker, GATK, and WDL in Terra , (O'Reilly Media, Sebastopol, CA, 2020). McLaren, W., et al. The Ensembl Variant Effect Predictor. Genome Biol 17, 122 (2016). Karczewski, K.J., et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581, 434–443 (2020). Amemiya, H.M., Kundaje, A. & Boyle, A.P. The ENCODE Blacklist: Identification of Problematic Regions of the Genome. Sci Rep 9, 9354 (2019). Additional Declarations Yes there is potential Competing Interest. Several authors are employees of F. Hoffmann-La Roche Ltd and/or Genentech, Inc., and may hold Roche stock or stock options as part of their employment compensation. Iosune Baraibar reports accommodation and travel expenses from Amgen, Merck, Sanofi, and Servier; and personal speaker honoraria from Astra Zeneca and Amgen. Victor Morena declared financial and non-financial interests with Abbvie, Achilles, Adaptimmune, Adc Therapeutics, Ascendis Pharma, Astrazeneca, Bayer, Beigene, Bicycle Tx, Bioinvent, Biomea Fusion, Biontech, Bms,, Boehringer, C4 Therapeutics, Calico Life Sciences Llc, Celgene, Constellation, Crescendo Biologics, Cullinan, Daiichi Sankyo, Debiopharm, Dragonfly, Enliven Therapeutics, Epizyme, Exelixis, Famewave, F-Star Beta Limited, Genentech, Genmab, Gilead, Grey Wolf Therapeutics, Gsk, Hexal Ag & Sandoz, Hifibio, Hookipa Biotech, Hutchmed, Igm Biosciences, Imcheck Therapeutics, Immunocore, Immutep, Incyte Iomx Therapeutics, Iovance, Italfarmaco, Iteos, Janssen, Light Chain Bioscience, Lilly, Loxo Oncology, Merck, Merus, Miltenyi Biomedicine, Monta Biosciences, Msd, Mythic Therapeutics, Ningbo Newbay, Novartis, Oxford Biotherapeutics, Pfizer, Pharmamar, Pmv Pharma, Prelude Therapeutics Inc, Pyxis Oncology, Regeneron, Relay Terapeutics, Repare Therapeutics, Revolution, Roche, Schrödinger, Scorpion Therapeutics, Seagen, Shattucks, Synthorx, Takeda, Tango Therapeutics, Tesaro, Totus Medicines, Turning Point Therapeutics, Vividion Therapeutics. Fiona Thistlethwaite declared institutional funding from Achilles Ltd, Adaptimmune, Amgen, Biontech, BMS, Byondis, Chugai, Corbus, Crescendo Biologics, Grey Wolf Therapeutics, GSK, Immunocore, Iovance, Janssen, Kymab Ltd/Sanofi, Leucid, Moderna, Novalgen, Nucana, Oxford Vacmedix, Roche, RS Oncology LLC, Seagen, Takeda, T-Knife Therapeutics, UCB, Zelluna, Zymeworks. Advisory board/consultancy/honoraria: AstraZeneca, Grey Wolf Therapeutics, OncoBayes, T-Knife Therapeutics, Waypoint DSMB: Immatics Ignacio Melero declared the following interests: Grants: Roche, AstraZenca, Bristol Myers, Genmab. Consultancy: Roche, Genmab, Regeneron, Pioneer medicines, Bright peak, Pharmamar, Curon, Mestag, Light chain, Emiliano Calvo reports the following interests: ownership interests in Oncoart Associated; honoraria from HM Hospitales Group; had consulting/advisory role at Nanobiotix, Janssen-Cilag, Roche/Genentech, TargImmune Therapeutics, Servier, Bristol-Myers Squibb, Amunix, Adcendo, Anaveon, AstraZeneca/MedImmune, Chugai Pharma, MonTa, MSD Oncology, Nouscom, Novartis, OncoDNA, T-Knife, Elevation Oncology, PharmaMar, Ellipses Pharma, Syneos Health, Genmab and Diaccurate; is president and founder of Foundation INTHEOS. Tae Won Kim declared institutional funding from Genentech and Inocras. Seung Hoon Beom declared the following interests: Researcher/Grant/Contract (clinical trial PI): Roche, MSD, Bayer, Boryung, Takeda, Jeil Pharmaceutical, Merck, Janssen, Pfizer, Amgen, Revolution Medicines, IGM Biosciences, STcube Supplementary Files SupplementaryInformation.docx Melero Tanos et al supplemental information Cite Share Download PDF Status: Published Journal Publication published 20 Apr, 2026 Read the published version in Nature Medicine → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8628656","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":577491959,"identity":"3f1b9c34-36df-40f5-b8d3-b6df27ecf0dd","order_by":0,"name":"Axel Boehnke","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIiWNgGAWjYHACxgOMDQwMBgxsDAwfGEAMIgBcC+MMkrUw8xCjhX9G8oEDjDts5Myl2xIf2+6xkTdnb2D88DEHtxaJM8cSDjCeSTO2nHPssHHOszTDnT0HmCVnbsNjzfEegwOMbYcTN9xIb5POOXA4weBGAhszLx4t8of5P4C01AO1tP+2AGm5/wC/FoPjPQwgLUDD044xM4BtYcCvxfDMMYMDiWfSDDfcSEuW7DkAZJxJbMbrF7kbyQ8ffNxhIw+0xfDDjwNAxvHDBz98xOd9EEhA5YKiaRSMglEwCkYBRQAAyj9ajkkldtwAAAAASUVORK5CYII=","orcid":"","institution":"F. Hoffmann-La Roche Ltd","correspondingAuthor":true,"prefix":"","firstName":"Axel","middleName":"","lastName":"Boehnke","suffix":""},{"id":577491960,"identity":"0837dd33-8b24-40c9-85d6-6663e5536f63","order_by":1,"name":"Ignacio Melero","email":"","orcid":"https://orcid.org/0000-0002-1360-348X","institution":"Clínica Universidad de Navarra Medical School, CIMA and CIBERONC","correspondingAuthor":false,"prefix":"","firstName":"Ignacio","middleName":"","lastName":"Melero","suffix":""},{"id":577491961,"identity":"7661c85e-f246-476b-a0aa-735b7e52be9b","order_by":2,"name":"Tamara Tanos","email":"","orcid":"","institution":"Roche","correspondingAuthor":false,"prefix":"","firstName":"Tamara","middleName":"","lastName":"Tanos","suffix":""},{"id":577491962,"identity":"93637923-c0c5-459c-9e29-6acac3299098","order_by":3,"name":"Emiliano Calvo","email":"","orcid":"https://orcid.org/0000-0003-4921-829X","institution":"START Madrid-CIOCC","correspondingAuthor":false,"prefix":"","firstName":"Emiliano","middleName":"","lastName":"Calvo","suffix":""},{"id":577491963,"identity":"6e632867-442d-45ee-b9d2-10d22dbe66d6","order_by":4,"name":"Camilla Qvortrup","email":"","orcid":"","institution":"Copenhagen University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Camilla","middleName":"","lastName":"Qvortrup","suffix":""},{"id":577491964,"identity":"089c2bc1-7c87-44b9-875c-a95efe5dfa08","order_by":5,"name":"Marloes van Dongen","email":"","orcid":"","institution":"The Netherlands Cancer Institute","correspondingAuthor":false,"prefix":"","firstName":"Marloes","middleName":"van","lastName":"Dongen","suffix":""},{"id":577491965,"identity":"0e531a0f-d9f7-4ff4-8377-f002f3f2f223","order_by":6,"name":"Iosune Baraibar","email":"","orcid":"https://orcid.org/0000-0002-3500-9641","institution":"Vall Hebron University Hospital and Vall Hebron Institute of Oncology (VHIO)","correspondingAuthor":false,"prefix":"","firstName":"Iosune","middleName":"","lastName":"Baraibar","suffix":""},{"id":577491966,"identity":"92fedd8e-e072-4191-b0ac-6af0bbb828b7","order_by":7,"name":"Seung-Hoon Beom","email":"","orcid":"","institution":"Yonsei Cancer Center, Yonsei University College of Medicine, Seoul, Republic of Korea","correspondingAuthor":false,"prefix":"","firstName":"Seung-Hoon","middleName":"","lastName":"Beom","suffix":""},{"id":577491967,"identity":"63b14afb-fad1-4912-83c1-bedd2f37d628","order_by":8,"name":"Fiona Thistlethwaite","email":"","orcid":"https://orcid.org/0000-0002-4832-7008","institution":"The Christie NHS Foundation Trust and University of Manchester","correspondingAuthor":false,"prefix":"","firstName":"Fiona","middleName":"","lastName":"Thistlethwaite","suffix":""},{"id":577491968,"identity":"73027881-5268-4542-b7b5-f0805d609260","order_by":9,"name":"Maria del Carmen Riesco","email":"","orcid":"","institution":"Medical Oncology Department, Hospital Universitario Doce de Octubre, Madrid, Spain","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"del Carmen","lastName":"Riesco","suffix":""},{"id":577491969,"identity":"d497e459-4033-482b-8110-7e6a9c936e95","order_by":10,"name":"Maria Martinez Garcia","email":"","orcid":"https://orcid.org/0000-0001-5891-3935","institution":"Medical Oncology Department, Hospital del Mar Research Institute, Barcelona, Spain and Centro de Investigación Biomédica en Red de Cáncer , Madrid, Spain","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"Martinez","lastName":"Garcia","suffix":""},{"id":577491970,"identity":"b09d172d-484d-4ceb-8ea1-f55149419c55","order_by":11,"name":"Victoria Woodcock","email":"","orcid":"","institution":"Oxford University Hospitals NHS Foundation Trust, Oxford, UK","correspondingAuthor":false,"prefix":"","firstName":"Victoria","middleName":"","lastName":"Woodcock","suffix":""},{"id":577491971,"identity":"c3b3ea35-e987-4f3f-ad5c-7378bb4d0af3","order_by":12,"name":"Tae Won Kim","email":"","orcid":"https://orcid.org/0000-0001-9522-1997","institution":"Asan Medical Center, University of Ulsan","correspondingAuthor":false,"prefix":"","firstName":"Tae","middleName":"Won","lastName":"Kim","suffix":""},{"id":577491972,"identity":"6e4d0c53-dc7c-4caa-be0f-f38d20a0b171","order_by":13,"name":"Pablo Umaña","email":"","orcid":"","institution":"Roche Innovation Center Zurich","correspondingAuthor":false,"prefix":"","firstName":"Pablo","middleName":"","lastName":"Umaña","suffix":""},{"id":577491973,"identity":"0de9f6f4-5500-445c-983c-a0946e2c531d","order_by":14,"name":"Christine McIntyre","email":"","orcid":"","institution":"Roche Pharma Research \u0026 Early Development, Roche Innovation Center Welwyn, Welwyn Garden City, UK","correspondingAuthor":false,"prefix":"","firstName":"Christine","middleName":"","lastName":"McIntyre","suffix":""},{"id":577491974,"identity":"8e4be399-ef22-4e03-acc9-91cef0be62e4","order_by":15,"name":"Lining Chen","email":"","orcid":"","institution":"Roche Pharma Research \u0026 Early Development, Roche Innovation Center Basel, Basel, Switzerland","correspondingAuthor":false,"prefix":"","firstName":"Lining","middleName":"","lastName":"Chen","suffix":""},{"id":577491975,"identity":"78021255-d0e5-4395-b528-970479a18303","order_by":16,"name":"Christian Heichinger","email":"","orcid":"","institution":"Roche Pharma Research \u0026 Early Development, Roche Innovation Center Basel, Basel, Switzerland","correspondingAuthor":false,"prefix":"","firstName":"Christian","middleName":"","lastName":"Heichinger","suffix":""},{"id":577491976,"identity":"38d037d4-41f1-4bc6-a4af-c739ef2823f4","order_by":17,"name":"Heather Hinton","email":"","orcid":"","institution":"F. Hoffmann-La Roche Ltd","correspondingAuthor":false,"prefix":"","firstName":"Heather","middleName":"","lastName":"Hinton","suffix":""},{"id":577491977,"identity":"9aeeef01-ea7a-4aa4-a423-ab033a8b2df2","order_by":18,"name":"Tulun Saylan","email":"","orcid":"","institution":"Roche Pharma Research \u0026 Early Development, Roche Innovation Center Basel, Basel, Switzerland","correspondingAuthor":false,"prefix":"","firstName":"Tulun","middleName":"","lastName":"Saylan","suffix":""},{"id":577491978,"identity":"781742cc-3dbe-4dd6-b2d8-53f2dfd2aebb","order_by":19,"name":"Iakov Davydov","email":"","orcid":"https://orcid.org/0000-0003-3510-3926","institution":"Computational Sciences Center of Excellence, Roche Innovation Center Basel, Basel, Switzerland","correspondingAuthor":false,"prefix":"","firstName":"Iakov","middleName":"","lastName":"Davydov","suffix":""},{"id":577491979,"identity":"26c7cff4-3a39-4fa9-aa7d-3fb0d93213c1","order_by":20,"name":"Ernesto Guarin","email":"","orcid":"","institution":"Roche Pharma Research \u0026 Early Development, Roche Innovation Center Basel, Basel, Switzerland","correspondingAuthor":false,"prefix":"","firstName":"Ernesto","middleName":"","lastName":"Guarin","suffix":""},{"id":577491980,"identity":"7311f668-8191-4679-b81e-601ed4951a07","order_by":21,"name":"Victor Moreno","email":"","orcid":"https://orcid.org/0000-0001-6099-4236","institution":"Hospital Fundación Jiménez Díaz","correspondingAuthor":false,"prefix":"","firstName":"Victor","middleName":"","lastName":"Moreno","suffix":""},{"id":577491981,"identity":"2e7b5c01-a5b3-4993-8233-68da964d13f0","order_by":22,"name":"not applicable cannot remove","email":"","orcid":"","institution":"F. Hoffmann-La Roche Ltd","correspondingAuthor":false,"prefix":"","firstName":"not","middleName":"applicable cannot","lastName":"remove","suffix":""}],"badges":[],"createdAt":"2026-01-18 01:20:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8628656/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8628656/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41591-026-04380-z","type":"published","date":"2026-04-20T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":102216709,"identity":"a4938535-2e6b-4931-8101-3703b6768757","added_by":"auto","created_at":"2026-02-09 12:58:16","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":375820,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpider plot – Change in tumor size for individual participants over time based on RECIST 1.1 (n=51)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSpider plot showing best percentage change from baseline in sum of longest diameters (SLD) by RECIST v1.1 versus study day. Lines are color-coded by FAP-4-1BBL regimen (QW, red; Q3W, blue). Each line represents an individual participant and reflects treatment duration and exposure (see Supplementary Table 7).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/ee8dfe940136718b23b0c738.png"},{"id":102216634,"identity":"329f5815-55d1-4b4d-b753-2883bb143ded","added_by":"auto","created_at":"2026-02-09 12:58:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":284782,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAntitumor activity results based on RECIST 1.1 and association with baseline characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWaterfall plot showing best percentage change from baseline in SLD by RECIST v1.1. Bars are color-coded by best overall response (BOR): PD (red), SD (yellow), and PR (blue). Heatmaps indicate baseline tumor and clinical features, including CEA and FAP expression, immune phenotype (inflamed, excluded, desert), liver metastases, prior lines of therapy, and dosing regimen. Gradients represent quantitative assessment for CEA and FAP H-score and CD8⁺/Ki67⁺ density.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/a6ba9943604e1bf777f0d658.png"},{"id":102216711,"identity":"634dbdce-0970-4722-bad0-4d62b5c83c71","added_by":"auto","created_at":"2026-02-09 12:58:16","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":895457,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBaseline biomarkers and response associations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBasal response association: (A) CEA IHC H-score in relation to ORR and DCR (n=44) and PFS stratified by H-score (\u0026lt;100, 101–200, ≥201–300; n=48). (B) Median FAP IHC intensity-scores versus ORR, DCR (n=43), and PFS by tertiles (n=47). (C) Distribution of consensus molecular subtypes (CMS1–4) among PR, SD, and PD (n=27), with PFS and OS by CMS category. (D) Best overall response by liver metastasis status (n=48) and PFS/OS according to liver metastasis presence (n=52). (E) Baseline immune phenotype distribution (inflamed, excluded, desert) by BOR (n=45), and corresponding associations with PFS and OS (n=47).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/8de861a33c76271521b3c849.png"},{"id":102216710,"identity":"60d167f8-8d90-4f63-ab41-e5568a4a5e3b","added_by":"auto","created_at":"2026-02-09 12:58:16","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":857324,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePeripheral pharmacodynamics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePeripheral blood pharmacodynamics: (A) Longitudinal cytokine log2 fold changes (pseudocount corrected) in participants treated with cibisatamab + FAP-4-1BBL (turquoise) or cibisatamab alone (red). Increases in IFNγ, TNFα, IL-10, and sustained sCD25 were observed, without additional IL-6 induction. (B) Peripheral immune cell dynamics. Heatmaps show fold-changes from baseline in CD8+ and CD4+ subsets across treatment, with contrasts between combination and monotherapy at C2 and D22. (C) Plasma sCD137 modulation by best overall response (BOR: PR, SD, PD). Increased sCD137 was seen across groups\u003csup\u003e26\u003c/sup\u003e, with greater and more sustained induction in PR. (D) Longitudinal sCEA levels by BOR. Declines were most pronounced in PR participants after Cycle 3, whereas minimal or no reduction was observed in PD.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/0cb445768ce0594925955400.png"},{"id":102216659,"identity":"b53ed1a3-31dd-46e3-af88-6a6cd43d1d2e","added_by":"auto","created_at":"2026-02-09 12:58:07","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1592116,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor pharmacodynamics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTumor Pharmacodynamics: (A) Representative duplex IHC of paired baseline and on-treatment MSS mCRC biopsies showing increased intratumoral CD8+ (yellow) and CD8+Ki67+ (purple) T cells after cibisatamab + FAP-4-1BBL. (B) Immune phenotype distribution, illustrating a shift from desert to inflamed/excluded phenotypes following cibisatamab +FAP-4-1BBL combination therapy. (C) Changes in CD8+ and CD8+Ki67+ T-cell density per participant for cibisatamab + FAP-4-1BBL combination (“combo”) versus cibisatamab monotherapy (“mono”), with fold-change distributions shown below. (D) RNA-seq analysis of paired biopsies (n=6) showing induction of CXCL10, IFNG, TNFRSF9 (4-1BB) and increased cytotoxic T-cell, IFNG (Rozeman21), and T-effector signatures. Lines denote individual participants by best overall response (PR, SD, PD).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/d3fcbaf70f3797a04e98362f.png"},{"id":102216618,"identity":"031f9836-4f11-42c5-82d1-558db0b78ae2","added_by":"auto","created_at":"2026-02-09 12:57:59","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":480429,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ectDNA/baseline and response associations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ectDNA response associations: (A) Baseline ctDNA levels by best overall response (BOR). (B) PFS and OS stratified by high (\u0026gt; median) versus low baseline ctDNA (≤ median). (C) ctDNA dynamics by BOR. Left: maximum fold-change to C3D1. Right: fold-change across timepoints. (D) PFS and OS by ctDNA fold-change from baseline to C3D1 (threshold: –0.81 log10), conditioned on survival to C3D1. Shaded areas denote 95% CI.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/8f8d15e670f7780c9a4f7d75.png"},{"id":107398152,"identity":"206bb334-ce8b-4fc8-8cf4-97ba192304fd","added_by":"auto","created_at":"2026-04-21 07:06:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5379221,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/b36f7532-1295-4e9f-b54a-9754e50c87e1.pdf"},{"id":102216638,"identity":"92667e5f-0765-42ba-8268-9d90d7c43875","added_by":"auto","created_at":"2026-02-09 12:58:03","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3430029,"visible":true,"origin":"","legend":"Melero Tanos et al supplemental information","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-8628656/v1/11f9a845130f7f1e522fda33.docx"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nSeveral authors are employees of F. Hoffmann-La Roche Ltd and/or Genentech, Inc., and may hold Roche stock or stock options as part of their employment compensation. \r\nIosune Baraibar reports accommodation and travel expenses from Amgen, Merck, Sanofi, and Servier; and personal speaker honoraria from Astra Zeneca and Amgen.\r\nVictor Morena declared financial and non-financial interests with Abbvie, Achilles, Adaptimmune, Adc Therapeutics, Ascendis Pharma, Astrazeneca, Bayer, Beigene, Bicycle Tx, Bioinvent, Biomea Fusion, Biontech, Bms,, Boehringer, C4 Therapeutics, Calico Life Sciences Llc, Celgene, Constellation, Crescendo Biologics, Cullinan, Daiichi Sankyo, Debiopharm, Dragonfly, Enliven Therapeutics, Epizyme, Exelixis, Famewave, F-Star Beta Limited, Genentech, Genmab, Gilead, Grey Wolf Therapeutics, Gsk, Hexal Ag \u0026 Sandoz, Hifibio, Hookipa Biotech, Hutchmed, Igm Biosciences, Imcheck Therapeutics, Immunocore, Immutep, Incyte Iomx Therapeutics, Iovance, Italfarmaco, Iteos, Janssen, Light Chain Bioscience, Lilly, Loxo Oncology, Merck, Merus, Miltenyi Biomedicine, Monta Biosciences, Msd, Mythic Therapeutics, Ningbo Newbay, Novartis, Oxford Biotherapeutics, Pfizer, Pharmamar, Pmv Pharma, Prelude Therapeutics Inc, Pyxis Oncology, Regeneron, Relay Terapeutics, Repare Therapeutics, Revolution, Roche, Schrödinger, Scorpion Therapeutics, Seagen, Shattucks, Synthorx, Takeda, Tango Therapeutics, Tesaro, Totus Medicines, Turning Point Therapeutics, Vividion Therapeutics.\r\nFiona Thistlethwaite declared institutional funding from Achilles Ltd, Adaptimmune, Amgen, Biontech, BMS, Byondis, Chugai, Corbus, Crescendo Biologics, Grey Wolf Therapeutics, GSK, Immunocore, Iovance, Janssen, Kymab Ltd/Sanofi, Leucid, Moderna, Novalgen, Nucana, Oxford Vacmedix, Roche, RS Oncology LLC, Seagen, Takeda, T-Knife Therapeutics, UCB, Zelluna, Zymeworks. Advisory board/consultancy/honoraria: AstraZeneca, Grey Wolf Therapeutics, OncoBayes, T-Knife Therapeutics, Waypoint DSMB: Immatics\r\nIgnacio Melero declared the following interests: Grants: Roche, AstraZenca, Bristol Myers, Genmab. Consultancy: Roche, Genmab, Regeneron, Pioneer medicines, Bright peak, Pharmamar, Curon, Mestag, Light chain, \r\nEmiliano Calvo reports the following interests: ownership interests in Oncoart Associated; honoraria from HM Hospitales Group; had consulting/advisory role at Nanobiotix, Janssen-Cilag, Roche/Genentech, TargImmune Therapeutics, Servier, Bristol-Myers Squibb, Amunix, Adcendo, Anaveon, AstraZeneca/MedImmune, Chugai Pharma, MonTa, MSD Oncology, Nouscom, Novartis, OncoDNA, T-Knife, Elevation Oncology, PharmaMar, Ellipses Pharma, Syneos Health, Genmab and Diaccurate; is president and founder of Foundation INTHEOS.\r\nTae Won Kim declared institutional funding from Genentech and Inocras.\r\nSeung Hoon Beom declared the following interests: Researcher/Grant/Contract (clinical trial PI): Roche, MSD, Bayer, Boryung, Takeda, Jeil Pharmaceutical, Merck, Janssen, Pfizer, Amgen, Revolution Medicines, IGM Biosciences, STcube","formattedTitle":"Tumor-targeted 4-1BB costimulation enhances immune activation and clinical activity of a CEA-directed T-cell engager in microsatellite-stable colorectal cancer","fulltext":[{"header":"Introduction","content":"\u003cp\u003eImmune checkpoint inhibitors have substantially improved outcomes in CRC with mismatch repair deficiency or high microsatellite instability (dMMR/MSI-H) \u003csup\u003e1\u003c/sup\u003e. However, more than 85% of patients present with mismatch repair\u0026ndash;proficient, microsatellite-stable (pMMR/MSS) disease \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, a largely non-inflamed tumor type in which immunotherapy has shown limited efficacy \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e, resulting in a persistently poor prognosis in the metastatic setting.\u003c/p\u003e \u003cp\u003eT-cell receptor\u0026ndash;engaging therapeutics can activate cytotoxic T lymphocytes and have demonstrated clinical benefit in hematologic malignancies \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e and, more recently, in selected solid tumors \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Nonetheless, extending these approaches to poorly immunogenic, non-inflamed tumors such as pMMR/MSS CRC remains challenging. Among inducible costimulatory receptors, 4-1BB (CD137) is upregulated following T-cell activation and promotes proliferation, survival, memory differentiation, and cytotoxic function upon engagement with its ligand or agonistic antibodies\u003csup\u003e\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eCibisatamab is a 2:1 T-cell bispecific antibody that bivalently binds a tumor-restricted carcinoembryonic antigen (CEA) epitope exposed on malignant cells after proteolytic cleavage, while engaging CD3ε on T cells. CEA is expressed on more than 80% of colorectal cancers on T cells \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. By simultaneously binding CEA and CD3ϵ, cibisatamab induces T-cell activation independently of native T-cell receptor specificity, resulting in lymphocyte-mediated tumor cell killing \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. In a prior phase 1 study of cibisatamab monotherapy in CEA-positive solid tumors, preliminary antitumor activity was observed, with confirmed partial responses in 4.0% of evaluable participants and a median duration of response of 6.5 months \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe hypothesized that the antitumor activity of a T-cell engager such as cibisatamab could be enhanced by spatially restricted costimulation mediated by a fibroblast activation protein (FAP)\u0026ndash;targeted 4-1BB ligand (FAP-4-1BBL). This concept is supported by preclinical coculture models of tumor organoids and FAP-expressing cancer-associated fibroblasts \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. FAP-4-1BBL enables selective 4-1BB engagement within FAP-expressing tissues, including the tumor microenvironment and tumor-draining lymph nodes \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. In a first-in-human dose-escalation study, FAP-4-1BBL monotherapy showed modest clinical activity but demonstrated an acceptable safety profile and evidence of increased intratumoral T-cell infiltration was observed \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eHere, we report the results of the phase 1b dose-escalation study BP42675 evaluating the combination of cibisatamab and FAP-4-1BBL following obinutuzumab pre-treatment in patients with heavily pretreated metastatic MSS CRC. Obinutuzumab pretreatment to deplete B cells was required because of almost constant induction of anti-drug antibodies (ADAs) by Cibisatamab\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. We further contextualize these findings in the combination setting by comparison with previously reported cibisatamab monotherapy data and associated pharmacodynamic analyses\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Our studies assessed the impact of tumor-localized 4-1BB costimulation on immune activation and clinical activity. Hereafter, RO7122290 is uniformly referred to as FAP-4-1BBL to ensure consistent terminology across the manuscript.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003eParticipants and Treatment\u003c/h2\u003e\n \u003cp\u003eFifty-two participants with pMMR/MSS mCRC who had progressed after at least two prior lines of therapy were enrolled in this open-label, multicenter, Phase 1b dose-escalation study. Baseline characteristics and demographics are summarized in Supplementary Data Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\n \u003cp\u003eThe study comprised two parts. Part 1 evaluated safety, pharmacokinetics (PK) and pharmacodynamics following escalation of FAP-4-1BBL with weekly dosing (QW). Doses considered safe in Part 1 were explored in Part 2 with FAP-4-1BBL administered every 3 weeks (Q3W) (Supplementary Figs. 1 and 2). In both parts, cibisatamab was administered Q3W.\u003c/p\u003e\n \u003cp\u003eUp to Part 1 Cohort 2 and Part 2 Cohort 1, cibisatamab was given at 100 mg from Cycle 1. Following a Grade 3 CRS event, the Cycle 1 cibisatamab dose was reduced to 60 mg from Part 1 Cohort 3 and Part 2 Cohort 2 onward, with 100 mg from Cycle 2 (cibisatamab 60\u0026rarr;100 mg). Cohort-level dosing details are provided in Baseline Characteristics Supplementary Table\u0026nbsp;1.\u003c/p\u003e\n \u003cp\u003eFAP-4-1BBL escalation in combination with cibisatamab was discontinued for cibisatamab-related strategic reasons (no safety signal observed) before a maximum tolerated dose (MTD) was defined.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eAnti-drug antibody management\u003c/h3\u003e\n\u003cp\u003ePrevious clinical experience with cibisatamab has demonstrated the induction of ADAs, primarily mediated by B-cell activation and differentiation into antibody-secreting plasma cells\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. To address this, the Phase 1b combination study of cibisatamab plus FAP-4-1BBL (BP42675) incorporated B-cell depletion with obinutuzumab as a pre-treatment, with a planned, repeated administration at six months. This strategy was designed to suppress de novo ADA formation while preserving pre-existing humoral immunity mediated by long-lived plasma cells, consistent with prior clinical observations\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. This study incorporated B-cell depletion with obinutuzumab as a pre-treatment, with a planned, repeat dosing at six months in patients remaining on treatment, to suppress ADA formation while preserving pre-existing humoral immunity mediated by long-lived CD20 negative plasma cells.\u003c/p\u003e\n\u003ch3\u003eSafety Outcomes\u003c/h3\u003e\n\u003cp\u003eThe safety profile of cibisatamab plus FAP-4-1BBL was manageable, with most adverse events (AEs) occurring early. Participants routinely received low-dose corticosteroid and non-steroidal anti-inflammatory premedication before infusions. The highest doses administered were 130 mg QW and 90 mg Q3W.\u003c/p\u003e\n\u003cp\u003eTwo DLTs were reported: presyncope (Grade 3 SAE, related to FAP-4-1BBL) in the FAP-4-1BBL QW 35 mg plus cibisatamab 100 mg cohort, and cytokine release syndrome (CRS; Grade 3 SAE, related to FAP-4-1BBL and cibisatamab) in the FAP-4-1BBL QW 50 mg plus cibisatamab 100 mg cohort. Overall safety is summarized in Supplementary Table\u0026nbsp;2.\u003c/p\u003e\n\u003cp\u003eAll participants (52/52) experienced at least one AE. The most common AEs (\u0026gt;\u0026thinsp;30%) were fatigue (33 participants, 63.5%), CRS (30, 57.7%), diarrhea (29, 55.8%), pyrexia (26, 50.0%), decreased appetite (23, 44.2%), cough (17, 32.7%), anemia (17, 32.7%), nausea (16, 30.8%) and arthralgia (16, 30.8%) (Supplementary Table\u0026nbsp;3). Grade\u0026thinsp;\u0026ge;\u0026thinsp;3 AEs occurred in 38 participants (73.1%), and SAEs in 34 participants (65.4%). Four participants (7.7%) had Grade 5 AEs (sepsis, sudden death, general physical health deterioration, and cytomegalovirus (CMV)-related colitis); the latter two were considered related to study treatment.\u003c/p\u003e\n\u003cp\u003eCRS occurred predominantly during Cycle 1 (median onset Day 2) and was Grade 1 in 25 participants (48.1%), Grade 2 in 3 (5.8%), and Grade 3 in 2 (3.8%). Serious CRS events occurred in 13 participants (25%) overall. In cohorts treated with cibisatamab 60 mg in Cycle 1 followed by 100 mg from Cycle 2, serious CRS occurred in 4/27 participants (14.8%). CRS events are summarized in Supplementary Table\u0026nbsp;4.\u003c/p\u003e\n\u003cp\u003eGastrointestinal toxicity, a known on-target/off-tumor effect of cibisatamab\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, was also observed with combination therapy: nausea (30.5%), vomiting (23.1%), and diarrhea (55.8%). Diarrhea occurred throughout treatment (median onset, 36 days; range, \u0026minus;\u0026thinsp;2 to 393 days relative to first dose). The negative onset reflects the study design, in which treatment was initiated on study day 1, while some participants received obinutuzumab lead-in dosing on study day\u0026thinsp;\u0026minus;\u0026thinsp;13 or \u0026minus;\u0026thinsp;8, depending on the protocol version. Colitis was reported in 7 participants (13.5%), including colitis, immune-mediated enterocolitis, and CMV colitis.\u003c/p\u003e\n\u003ch3\u003ePharmacokinetics and Immunogenicity Outcomes\u003c/h3\u003e\n\u003cp\u003eFAP-4-1BBL exposure increased with dose. Geometric mean PK parameters, including AUC and Cmax, rose with escalating doses. Elimination was non-linear, consistent with target-mediated drug disposition (TMDD)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. PK parameters are summarized in Supplementary Table\u0026nbsp;5.\u003c/p\u003e\n\u003cp\u003eCibisatamab exposure was consistent with prior reports in monotherapy and in combination with atezolizumab, with Cycle 1 Cmax and AUC within established ranges\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, suggesting no meaningful impact of FAP-4-1BBL co-administration on cibisatamab PK. Cycle 1 PK parameters are summarized in Supplementary Table\u0026nbsp;6.\u003c/p\u003e\n\u003cp\u003eTwo participants pretreated with obinutuzumab developed ADAs to FAP-4-1BBL; no participants had treatment-induced ADAs to cibisatamab.\u003c/p\u003e\n\u003ch3\u003eEfficacy Outcomes\u003c/h3\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003eResponse Rates\u003c/h2\u003e\n \u003cp\u003eFifty-one of 52 participants were evaluable for response (one participant had no on-treatment tumor assessment owing to clinical deterioration). Partial responses (PR) by RECIST 1.1 were observed in 9 participants (17.6%, 95% CI: 6.06\u0026ndash;28.55), including 7 confirmed PRs (13.7%, 95% CI: 3.22\u0026ndash;23.70). No complete responses were observed.\u003c/p\u003e\n \u003cp\u003ePRs occurred in 7/29 participants (24.1%) receiving FAP-4-1BBL QW and 2/22 (9.1%) receiving FAP-4-1BBL Q3W. By cibisatamab dosing strategy, PRs occurred in 4/25 participants (16.0%) treated with 100 mg from Cycle 1 and 5/26 (19.2%) treated with cibisatamab 60\u0026rarr;100 mg (step-up). In the highest targeted QW dose groups (90 mg and 130 mg combined; n\u0026thinsp;=\u0026thinsp;11; all treated with cibisatamab 60\u0026rarr;100 mg), PRs occurred in 4 participants (36.4%), including 3 confirmed PRs (27.3%). In the highest targeted Q3W dose group (90 mg; n\u0026thinsp;=\u0026thinsp;3; all treated with cibisatamab 60\u0026rarr;100 mg), no PRs were observed (limited enrollment due to strategic considerations).\u003c/p\u003e\n \u003cp\u003eThe disease control rate (DCR) was 50.9% (26/51), including 58.6% (17/29) with FAP-4-1BBL QW and 40.9% (9/22) with FAP-4-1BBL Q3W. DCR was 56.0% (14/25) in participants treated with cibisatamab 100 mg from Cycle 1 and 46.1% (12/26) in participants treated with cibisatamab 60\u0026rarr;100 mg, irrespective of FAP-4-1BBL schedule. In the highest targeted QW dose groups (90 mg and 130 mg combined; n\u0026thinsp;=\u0026thinsp;11), DCR was 63.6%.\u003c/p\u003e\n \u003cp\u003eCohort-level efficacy is summarized in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e, categorized by FAP-4-1BBL dose and schedule and by cibisatamab dosing strategy (cibisatamab 100 mg versus cibisatamab 60\u0026rarr;100 mg).\u003c/p\u003e\n \u003cp\u003eProgression-free survival (PFS) by investigator assessment suggested a trend toward longer disease control in participants treated in the later dose-escalation cohorts Part 1 Cohort 4 (P1C4) and Part 1 Cohort 5 (P1C5) (n\u0026thinsp;=\u0026thinsp;12), compared with earlier Part 1 Cohorts 1\u0026ndash;3 (P1C1\u0026ndash;P1C3) (n\u0026thinsp;=\u0026thinsp;18) (Supplementary Fig.\u0026nbsp;3). Part 1 Cohorts 1\u0026ndash;3 evaluated earlier FAP-4-1BBL dose levels with cibisatamab administered at a flat 100-mg dose starting from Cycle 1, whereas Part 1 Cohorts 4 and 5 explored higher and optimized FAP-4-1BBL dose levels in combination with a refined cibisatamab step-up regimen (60 mg at Cycle 1 followed by 100 mg from Cycle 2). Although progression events occurred in both groups, separation of the PFS curves became apparent at later timepoints, with a higher proportion of participants in Part 1 Cohorts 4\u0026ndash;5 remaining progression-free beyond approximately 150 days. Median PFS was not reached in the Part 1 Cohort 4\u0026ndash;5 group at the time of analysis, whereas participants in earlier Part 1 cohorts experienced progression earlier. These findings are exploratory and should be cautiously interpreted given the limited sample size and non-randomized cohort structure.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eTable 1. Best overall response by investigator (RECIST v1.1)\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eCibisatamab + FAP-4-1BBL. Confirmed and unconfirmed responses (left) and confirmed responses only (right).\u003c/p\u003e\n \u003cdiv\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"634\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCombination schedule\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFAP-4-1BBL dose (mg)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCibisatamab dose (mg)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e\u003cstrong\u003en evaluable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePR n (%)\u003cbr\u003e\u0026nbsp;(95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSD n (%)\u003cbr\u003e\u0026nbsp;(95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eORR n (%)\u003cbr\u003e\u0026nbsp;(95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDCR n (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConfirmed PR n (%)\u003cbr\u003e\u0026nbsp;(95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConfirmed SD n (%)\u003cbr\u003e\u0026nbsp;(95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConfirmed ORR n (%)\u003cbr\u003e\u0026nbsp;(95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConfirmed DCR n (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFAP-4-1BBL QW + Cibisatamab Q3W\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e12**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e2 (16.7)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;41.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e5 (41.7)\u003cbr\u003e\u0026nbsp;(9.61\u0026ndash;73.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e2 (16.7)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;41.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e7 (58.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e2 (16.7)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;41.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e5 (41.7)\u003cbr\u003e\u0026nbsp;(9.61\u0026ndash;73.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e2 (16.7)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;41.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e7 (58.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e1 (33.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e1 (33.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e60\u0026rarr;100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e2 (66.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e2 (66.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e60\u0026rarr;100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e6*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e3 (50.0)\u003cbr\u003e\u0026nbsp;(1.66\u0026ndash;98.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (16.7)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;54.82)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e3 (50.0)\u003cbr\u003e\u0026nbsp;(1.66\u0026ndash;98.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e4 (66.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e2 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;79.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e2 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;79.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e2 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;79.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e4 (66.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e1 (20.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;65.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e2 (40.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;92.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (20.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;65.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e3 (60.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (20.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;65.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e2 (40.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;92.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e1 (20.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;65.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e3 (60.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFAP-4-1BBL Q3W + Cibisatamab Q3W\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e2 (20.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;49.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e4 (40.0)\u003cbr\u003e\u0026nbsp;(4.64\u0026ndash;75.36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e2 (20.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;49.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e6 (60.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (10.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;33.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e5 (50.0)\u003cbr\u003e\u0026nbsp;(14.01\u0026ndash;85.99)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e1 (10.0)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;33.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e6 (60.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e60\u0026rarr;100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;5.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e2 (22.2)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;54.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;5.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e2 (22.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;5.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e2 (22.2)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;54.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;5.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e2 (22.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003e90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003e60\u0026rarr;100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e1 (33.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e1 (33.3)\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e0\u003cbr\u003e\u0026nbsp;(0.00\u0026ndash;16.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e1 (33.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 10.2524%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 5.83596%;\"\u003e\n \u003cp\u003eAll\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 6.46688%;\"\u003e\n \u003cp\u003eAll\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 4.73186%;\"\u003e\n \u003cp\u003e51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 12.3028%;\"\u003e\n \u003cp\u003e9 (17.6)\u003cbr\u003e\u0026nbsp;(6.20\u0026ndash;29.09)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e17 (33.3)\u003cbr\u003e\u0026nbsp;(19.42\u0026ndash;47.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e9 (17.6)\u003cbr\u003e\u0026nbsp;(6.20\u0026ndash;29.09)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.57098%;\"\u003e\n \u003cp\u003e26 (50.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.46372%;\"\u003e\n \u003cp\u003e7 (13.7)\u003cbr\u003e\u0026nbsp;(3.30\u0026ndash;24.15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e19 (37.3)\u003cbr\u003e\u0026nbsp;(23.01\u0026ndash;51.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 8.51735%;\"\u003e\n \u003cp\u003e7 (13.7)\u003cbr\u003e\u0026nbsp;(3.30\u0026ndash;24.15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.41325%;\"\u003e\n \u003cp\u003e26 (50.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003eAbbreviations: PR, partial response; SD, stable disease; ORR, objective response rate; DCR, disease control rate; QW, once weekly; Q3W, once every 3 weeks. Cibisatamab 60\u0026rarr;100 indicates 60 mg at Cycle 1 followed by 100 mg from Cycle 2 onward.\u003c/p\u003e\n \u003cp\u003e* Seven participants were enrolled in this cohort; one withdrew due to an adverse event and had no RECIST assessment.\u003c/p\u003e\n \u003cp\u003e** One participant withdrew after a single study dose (FAP-4-1BBL + cibisatamab) but was later reported as a confirmed PR without receiving any anticancer therapy, per investigator assessment.\u003c/p\u003e\n \u003cp\u003eTreatment duration suggested longer exposure with QW versus Q3W FAP-4-1BBL (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). For example, median FAP-4-1BBL treatment duration was 123 days (range 1\u0026ndash;435) in the 35 mg QW cohort plus cibisatamab 100 mg (n\u0026thinsp;=\u0026thinsp;12) versus 64 days (range 1\u0026ndash;260) in the corresponding 35 mg Q3W cohort (n\u0026thinsp;=\u0026thinsp;10). Similarly, in the 90 mg cohorts, median duration was 99 days (range 1\u0026ndash;232) for QW (n\u0026thinsp;=\u0026thinsp;6) versus 43 days (range 22\u0026ndash;57) for Q3W (n\u0026thinsp;=\u0026thinsp;3). In cohorts treated with the same FAP-4-1BBL dose and schedule, longer treatment duration was observed in participants who received cibisatamab 60\u0026rarr;100 mg versus 100 mg from Cycle 1 (Supplementary Table\u0026nbsp;7).\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eExposure-Safety Analysis\u003c/h3\u003e\n\u003cp\u003eWe evaluated associations between FAP-4-1BBL exposure and AEs of interest (CRS, infusion-related reactions (IRR), rash, diarrhea, and colitis). For acute events (CRS and IRR), analyses used first-dose Cmax (evaluable: QW n\u0026thinsp;=\u0026thinsp;29, Q3W n\u0026thinsp;=\u0026thinsp;22). For delayed events (rash, diarrhea, colitis), analyses used cumulative exposure over the first 63 days (AUC0\u0026ndash;63 days) to approximate exposure through approach to steady state.\u003c/p\u003e\n\u003cp\u003eNo consistent exposure\u0026ndash;AE relationships were observed for CRS, IRR, rash, or colitis across regimens. A possible trend between higher exposure and diarrhea was noted in the QW regimen but not in Q3W. Owing to low event counts, logistic regression analyses were underpowered for definitive inference. Additional details are provided in Supplementary Fig.\u0026nbsp;4.\u003c/p\u003e\n\u003ch3\u003eAnalysis of Baseline Markers in Association with Clinical Response\u003c/h3\u003e\n\u003cp\u003eTo explore predictors of response, we assessed best change in SLD across participants (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). PRs occurred across a range of baseline biomarker profiles, with clinical benefit not restricted to a single subgroup. Heatmap overlays of CEA and FAP expression, immune phenotype, prior therapy, and baseline tumor CD8⁺ infiltration suggested responses also occurred in participants with adverse features, including liver metastases and excluded/desert immune phenotypes.\u003c/p\u003e\n\u003cp\u003eBaseline profiling identified tumor-intrinsic and clinical features associated with outcome. Tumor CEA by IHC was associated with DCR (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eA): participants with low CEA (H-score\u0026thinsp;\u0026le;\u0026thinsp;100) had shorter PFS than those with medium (101\u0026ndash;200) or high (201\u0026ndash;300) scores. This association was not reproduced when CEACAM5 was assessed by RNA-seq (Supplementary Fig.\u0026nbsp;5).\u003c/p\u003e\n\u003cp\u003eBaseline FAP expression did not correlate with DCR or ORR; tertile stratification suggested a favorable PFS trend in the intermediate-expression group (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eB).\u003c/p\u003e\n\u003cp\u003eWaterfall plot showing best percentage change from baseline in SLD by RECIST v1.1. Bars are color-coded by best overall response (BOR): PD (red), SD (yellow), and PR (blue). Heatmaps indicate baseline tumor and clinical features, including CEA and FAP expression, immune phenotype (inflamed, excluded, desert), liver metastases, prior lines of therapy, and dosing regimen. Gradients represent quantitative assessment for CEA and FAP H-score and CD8⁺/Ki67⁺ density.\u003c/p\u003e\n\u003cp\u003eConsensus molecular subtype (CMS) analysis \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e suggested enrichment of responders in CMS2, an epithelial, WNT/MYC-driven transcriptional subtype (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eC). Although the overall association between CMS category and clinical outcome did not reach statistical significance (P\u0026thinsp;=\u0026thinsp;0.069), participants with CMS2 tumors showed numerical improvements in both PFS and OS. No partial responses (PRs) were observed in CMS1 or CMS4 tumors. CMS1 tumors, which are enriched for immune activation, interferon signaling, and MSI features and are frequently associated with a more favorable prognosis in early-stage disease, did not derive objective benefit in this heavily pretreated pMMR/MSS metastatic cohort. CMS4 tumors, characterized by epithelial\u0026ndash;mesenchymal transition and a TGF-\u0026beta;\u0026ndash;driven immunosuppressive phenotype, were likewise non-responsive \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003ePretreatment immune phenotyping showed that most partial responses (PRs) occurred in tumors with an excluded immune phenotype, whereas progressive disease (PD) was more frequently observed in immune-desert tumors.\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. The T-cell excluded phenotype was associated with longer PFS (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eE). In contrast, neither baseline tumor 4-1BB⁺ or CD8⁺ density nor circulating cytokines/immune-cell subsets (including IL-8, lymphocyte-to-neutrophil ratio, circulating CD8⁺/Ki67⁺ T cells, IFN-\u0026gamma;, CXCL10, soluble IL-2 receptor, and soluble target proteins) were associated with clinical outcomes (Supplementary Fig.\u0026nbsp;5).\u003c/p\u003e\n\u003cp\u003eActive liver metastases, an adverse prognostic factor\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, did not influence response, PFS, or OS; 4 of 7 of the confirmed PRs (57%) occurred in participants with at least one liver target lesion (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eD).\u003c/p\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003ePharmacodynamics\u003c/h2\u003e\n \u003cp\u003ePharmacodynamic effects were assessed in peripheral blood and paired tumor biopsies.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003ePharmacodynamic changes upon treatment in peripheral blood\u003c/h2\u003e\n \u003cp\u003ePeripheral blood cytokines and immune-cell populations were analyzed and compared with published cibisatamab monotherapy data \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, and further extended by additional pharmacodynamic analyses performed on previously unpublished data from that study.\u003c/p\u003e\n \u003cp\u003eCibisatamab plus FAP-4-1BBL induced increases in circulating IFN\u0026gamma;, soluble CD25 (sCD25), and IL-6 (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). IFN\u0026gamma; rose by Day 15 and remained elevated; sCD25 showed a similar sustained pattern, with both markers induced more strongly than with cibisatamab alone. IL-6 increased with kinetics similar to monotherapy. TNF\u0026alpha; and IL-10 were also higher at Day 15, consistent with an enhanced immunostimulatory profile.\u003c/p\u003e\n \u003cp\u003eFlow cytometry showed an early, transient circulating lymphocyte reduction (8 hours and C1D2), consistent with margination and/or trafficking, followed by return to baseline.\u003c/p\u003e\n \u003cp\u003eThe combination increased proliferating CD8⁺Ki67⁺ T cells and activated CD8⁺ subsets (Tim3, 4-1BB, HLA-DR), with expansion of central and effector memory compartments. Compared with cibisatamab monotherapy, Day 22 signs of activation were greater, including higher CD8⁺4-1BB⁺ and CD8⁺HLA-DR⁺ frequencies and increased PD-1 (CD279) expression on central memory and effector CD8⁺ T cells, consistent with more pronounced CD8⁺ activation.\u003c/p\u003e\n \u003cp\u003eSoluble 4-1BB (sCD137) increased across response groups, consistent with target engagement\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. Participants with PR showed more sustained induction during cycles 1\u0026ndash;2, whereas SD/PD showed lower and more variable changes.\u003c/p\u003e\n \u003cp\u003eSerum CEA (sCEA) declined in participants with PR from Cycle 2 onward (median\u0026thinsp;~\u0026thinsp;0.5\u0026ndash;1.0 log reductions by cycles 6\u0026ndash;8), with smaller changes in SD and little change in PD.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003ePharmacodynamic changes upon treatment in fresh paired tumor biopsies\u003c/h2\u003e\n \u003cp\u003ePaired fresh biopsies were collected at baseline and Day 23 after the first combination dose to assess intratumoral immune pharmacodynamics.\u003c/p\u003e\n \u003cp\u003eDuplex IHC (CD8, Ki67) showed\u0026thinsp;\u0026gt;\u0026thinsp;2-fold increases in intratumoral CD8⁺ T cells in 12/14 evaluable pairs and \u0026gt;\u0026thinsp;2-fold increases in CD8⁺Ki67⁺ cells in 10/14 pairs (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eA; Supplementary Fig.\u0026nbsp;6). Immune phenotype profiling\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e classified baseline lesions as desert (79%), excluded (14%) or inflamed (7%). On treatment, desert lesions decreased to 21%, with on-treatment assessed increases in excluded (50%) and inflamed (29%) phenotypes (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eB).\u003c/p\u003e\n \u003cp\u003eQuantitative image analysis showed higher post-treatment CD8⁺ and CD8⁺Ki67⁺ densities after combination therapy than after cibisatamab monotherapy: median intratumoral CD8⁺ cell density was approximately threefold higher than that observed with monotherapy (p\u0026thinsp;=\u0026thinsp;0.056), with a corresponding fivefold increase in CD8⁺Ki67⁺ cell density (p\u0026thinsp;=\u0026thinsp;0.019) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ec). Such increases were also observed in 10/11 participants with liver metastases (Supplementary Fig.\u0026nbsp;6).\u003c/p\u003e\n \u003cp\u003eRNA-seq of paired biopsies passing QC (n\u0026thinsp;=\u0026thinsp;6) showed induction of immune effector genes (CXCL9, CXCL10, IFN\u0026gamma;, TNFRSF9/4-1BB) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eD; Supplementary Fig.\u0026nbsp;6). IFN-\u0026gamma; response, cytotoxic T-cell, T-effector and antigen processing machinery (APM) signatures increased (all p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). Induction was greater in PR/SD than PD despite higher baseline expression in PD, consistent with a less robust transcriptional response in PD tumors.\u003c/p\u003e\n \u003cp\u003eA modest but significant increase in the M1/M2 macrophage ratio was estimated based on gene signatures (Supplementary Fig.\u0026nbsp;6). Adaptive immune-regulatory programs were also induced, including PD-L1/CD274, LAG3, EOMES\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e27\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e and IDO1\u003csup\u003e29,30\u003c/sup\u003e (Supplementary Fig.\u0026nbsp;6). FAP and CEA expression showed no consistent treatment-induced change (Supplementary Fig.\u0026nbsp;6).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003eCirculating ctDNA\u003c/h2\u003e\n \u003cp\u003eIn an exploratory tissue-informed patient-specific multiplex PCR assay, baseline ctDNA levels varied widely (median 1,604 MTM/ml) and were generally lower in participants with PR (with one exception) than in participants with SD or PD (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eA). Lower baseline ctDNA showed a non-significant trend toward improved PFS (p\u0026thinsp;=\u0026thinsp;0.061) and OS (p\u0026thinsp;=\u0026thinsp;0.2) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eB).\u003c/p\u003e\n \u003cp\u003eBy Cycle 3 Day 1 (C3D1), lower residual ctDNA (\u0026lt;\u0026thinsp;384.42 MTM/ml) was associated with longer PFS (p\u0026thinsp;=\u0026thinsp;0.008) and OS (p\u0026thinsp;=\u0026thinsp;0.033) (Supplementary Fig.\u0026nbsp;7). From baseline to C3D1, ctDNA decreased\u0026thinsp;\u0026gt;\u0026thinsp;3-fold in 5/5 participants with PR and 6/8 with SD, whereas participants with PD showed minimal change or increases (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eC). Greater ctDNA clearance was associated with longer PFS (p\u0026thinsp;=\u0026thinsp;0.009) (Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e6\u003c/span\u003eD).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study provides clinical evidence in solid tumors that coordinated delivery of synthetic signal 1 and signal 2\u003csup\u003e31,32\u003c/sup\u003e can be achieved through the combination of a tumor-directed T-cell engager and a spatially restricted costimulatory agonist. As a signal‑1 provider, the CEA-directed T-cell engager cibisatamab induces antigen-dependent T-cell activation and tumor infiltration \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e and has previously demonstrated modest but reproducible clinical activity in heavily pretreated MSS colorectal cancer patients at doses comparable to those evaluated here \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOptimal T-cell activation requires integration of TCR\u0026ndash;CD3 signaling with costimulatory inputs, including those mediated by 4‑1BB and other members of the TNFR family\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Importantly, 4-1BB is selectively upregulated following initial T-cell priming and is undetectable on resting T lymphocytes\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Based on this biology, we hypothesized that coupling cibisatamab-mediated T-cell engagement with FAP‑4‑1BBL\u0026ndash;mediated costimulation would amplify antitumor immunity while restricting signal‑2 delivery to the tumor microenvironment, where FAP is selectively expressed by cancer-associated fibroblasts outside contexts of acute tissue repair or scar formation\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn the BP42675 study, the combination of cibisatamab and FAP‑4‑1BBL following obinutuzumab pre-treatment demonstrated a manageable safety profile, robust pharmacodynamic immune activation, and encouraging preliminary efficacy in patients with heavily pretreated MSS mCRC. Obinutuzumab was incorporated to mitigate anti-drug antibody formation previously observed with cibisatamab monotherapy, thereby enabling sustained drug exposure and preservation of biological activity \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Although B-cell depletion raises theoretical concerns regarding disruption of tertiary lymphoid structures and humoral support of antitumor immunity\u003csup\u003e\u003cspan additionalcitationids=\"CR37\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e, no clear detriment to clinical or immunologic outcomes was observed in this study. Nevertheless, the absence of a non\u0026ndash;B-cell-depleted comparator cohort represents an important limitation and warrants further evaluation since pro-tumor activities of B lymphocytes have also been reported\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDose and schedule selection for cibisatamab was informed by prior clinical experience and by preclinical imaging studies demonstrating prolonged intratumoral retention despite a short systemic half-life \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. For FAP‑4‑1BBL, both weekly and every‑3‑week dosing regimens were explored to evaluate the impact of sustained versus intermittent costimulatory signaling, an area of ongoing debate for agonist 4‑1BB biology\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Notably, numerically higher response rates, deeper tumor shrinkage, and longer treatment durations were observed in cohorts receiving weekly FAP‑4‑1BBL, particularly at higher dose levels, suggesting that more continuous costimulatory engagement may be advantageous in this setting.\u003c/p\u003e \u003cp\u003eAcross the response-evaluable population, cibisatamab plus FAP‑4‑1BBL achieved numerically higher ORR and DCR than previously reported for cibisatamab monotherapy \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, with the greatest activity observed in higher-dose FAP‑4‑1BBL weekly cohorts. These findings are consistent with preclinical evidence demonstrating synergistic antitumor effects of T-cell engagement combined with 4‑1BB costimulation in murine models and patient-derived tumor systems \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. While cross-trial comparisons should be interpreted cautiously, the convergence of clinical and pharmacodynamic signals supports a true biological interaction between signal‑1 and signal‑2 to achieve a more pronounced T-cell activation and tumor infiltration.\u003c/p\u003e \u003cp\u003eThe safety profile of the combination was largely consistent with the known toxicities of cibisatamab and with prior experience from the FAP‑4‑1BBL first-in-human study. Gastrointestinal toxicity, a known on-target/off-tumor effect of CEA-directed T-cell engagement \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, was common but generally manageable and not dose-limiting. Importantly, the addition of tumor-targeted 4‑1BB costimulation did not result in excess systemic immune toxicity, contrasting with historical experience using untargeted 4‑1BB agonist antibodies \u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eExtensive biomarker analyses provided mechanistic context for treatment activity. Analyses were primarily contextualized against published and additional cibisatamab monotherapy data in patients with mCRC \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, as this represents the most relevant disease-matched comparator. In contrast, the first-in-human FAP-4-1BBL monotherapy study enrolled multiple tumor types, limiting its utility for direct comparison in MSS mCRC\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. Nevertheless, integration of data across all three studies indicates that the cibisatamab plus FAP-4-1BBL combination induced the largest increases in intratumoral CD8⁺ and proliferating CD8⁺Ki67⁺ T cells, accompanied by concordant peripheral immune activation \u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. In BP42675, these effects were observed even in liver metastases, suggesting partial mitigation of liver-associated immune suppression.\u003c/p\u003e \u003cp\u003eBaseline biomarker analyses identified tumor CEA protein expression by immunohistochemistry as a potential predictor of disease control and progression-free survival. Intermediate FAP expression was associated with more favorable outcomes, suggesting that effective costimulation may require sufficient\u0026mdash;but not excessive\u0026mdash;CAF targeting \u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. Responses were enriched in CMS2 tumors and in lesions with an immune-excluded phenotype, indicating that tumors traditionally considered poorly inflamed may still be amenable to T-cell engager\u0026ndash;based strategies when combined with tumor-localized artificial costimulation\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e,\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOn-treatment biomarkers further underscored the depth of immune activation achieved with the combination, including sustained induction of IFNγ, soluble CD25, activated CD8⁺ T-cell subsets, and soluble CD137, consistent with engagement of the 4-1BB pathway\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. A transient early decrease in circulating lymphocyte counts was observed within hours of dosing, consistent with lymphocyte margination and rapid immune-cell redistribution, a phenomenon previously described for T-cell engager therapies\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e,\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. In addition, early reductions in serum CEA and circulating tumor DNA (ctDNA) were associated with improved clinical outcomes, supporting their utility as minimally invasive pharmacodynamic and early response biomarkers \u003csup\u003e\u003cspan additionalcitationids=\"CR51\" citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn the context of robust and sustained immune activation, treatment was accompanied by increased expression of several inhibitory checkpoint molecules and metabolic regulators, including PD-L1, LAG3, EOMES, and IDO1, consistent with adaptive feedback mechanisms that can emerge under heightened immune stimulation \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e. Notably, these transcriptional changes occurred alongside a treatment-associated shift toward a more pro-inflammatory myeloid milieu, reflected, for example, by an increased M1 to M2 macrophage polarization signature ratio and upregulation of antigen presentation machinery signatures, which was not observed with cibisatamab monotherapy\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e, and accompanied by marked CD8⁺ T-cell infiltration and activation. Together, these findings indicate that immune activation predominates over counter-regulatory signals and support the biological activity of the combination, while suggesting that simultaneous or sequential co-treatment with PD-(L)1 checkpoint inhibitors may represent a rational strategy to enhance response durability \u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn conclusion, this study establishes a clinically tractable framework for coordinated delivery of signal-1 and signal-2 immunostimulation within the tumor microenvironment of solid tumors. By pairing a tumor-directed T-cell engager with spatially restricted 4-1BB costimulation, cibisatamab plus FAP-4-1BBL achieved enhanced intratumoral T-cell activation, encouraging antitumor activity, and a manageable safety profile in a disease historically refractory to immunotherapy. These data provide proof of principle that localized costimulation can amplify the efficacy of T-cell engagers without recapitulating the systemic toxicities associated with untargeted agonists. Beyond MSS colorectal cancer, this strategy offers a modular and broadly applicable paradigm for overcoming immune exclusion and resistance in non-inflamed tumors and supports further clinical development of T-cell engager\u0026ndash;based combinations incorporating tumor-restricted costimulatory signals to improve the depth and durability of response.\u003c/p\u003e "},{"header":"Methods","content":"\u003cp\u003e\u003cu\u003eInclusion and ethics\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eThis study was an open-label, multicenter, phase 1b dose-escalation trial conducted across 15 globally approved clinical sites, of which 12 were actively recruiting participants. The trial was performed in accordance with Good Clinical Practice guidelines and applicable regulatory requirements. Approval was obtained from institutional review boards or independent ethics committees at each participating site. All participants provided written informed consent prior to enrolment and before the initiation of any study-related procedures.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eTrial design, participants and treatments\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eThe trial consisted of two sequential dose-escalation parts designed to explore different FAP-4-1BBL exposure profiles. Part 1 evaluated continuous exposure to FAP-4-1BBL administered weekly (QW), whereas Part 2 investigated a pulsatile exposure strategy with FAP-4-1BBL administered every 3 weeks (Q3W In both study parts, cibisatamab was administered intravenously Q3W, with dosing consisting of either 100 mg in all cycles or a step-up regimen of 60 mg in cycle 1 followed by 100 mg thereafter. (Supplementary Figures 1 and 2).\u003c/p\u003e\n\u003cp\u003eAt study initiation, participant eligibility required high tumor CEACAM5 expression determined by quantitative PCR analysis of FFPE tumor tissue, consistent with prior cibisatamab studies \u003csup\u003e16\u003c/sup\u003e. Following enrollment of the first 11 participants, this requirement was removed to enable assessment of the cibisatamab and FAP-4-1BBL combination in an unselected colorectal cancer population, reflecting the high prevalence of CEA expression in CRC \u003csup\u003e16\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTo mitigate the development of anti-drug antibodies (ADAs) against cibisatamab, as previously reported \u003csup\u003e16\u003c/sup\u003e, all participants received B-cell\u0026ndash;depleting obinutuzumab as pre-treatment 1\u0026ndash;2 weeks prior to Cycle 1 Day 1. Obinutuzumab dosing was repeated after 6 months of treatment.\u003c/p\u003e\n\u003cp\u003eAdverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 5.0. Cytokine release syndrome (CRS) was graded according to the American Society for Transplantation and Cellular Therapy consensus criteria. Study objectives and corresponding endpoints for BP42675 are summarized in Supplementary Table 8.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eStatistical analysis\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eAll 52 enrolled participants were evaluable for safety and dose-limiting toxicities (DLTs). One participant did not undergo an on-treatment tumor assessment and was therefore excluded from efficacy analyses, resulting in an efficacy-evaluable population of 51 participants. All analyses were exploratory in nature.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003ePK sampling and parameter calculation\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eSerial pharmacokinetic (PK) samples were collected from all participants to characterize the PK profiles of FAP-4-1BBL and cibisatamab following initial administration. Based on these data, key PK exposure parameters were calculated for each participant using standard non-compartmental analysis (NCA). Only participants with a sufficient number of PK samples to allow robust NCA were included in exposure\u0026ndash;response analyses.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eExposure\u0026ndash;safety analysis\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003ePotential associations between FAP-4-1BBL exposure and the incidence and severity of adverse events of particular interest were evaluated in all safety-evaluable participants. Adverse events assessed included CRS, infusion-related reactions (IRR), rash, diarrhea, and colitis.\u003c/p\u003e\n\u003cp\u003eThe PK exposure metric selected for analysis was tailored to the anticipated timing and mechanism of each adverse event. For acute events such as CRS and IRR, which typically occur shortly after infusion, exposure was evaluated using maximum plasma concentration (C_max) following the first dose. For adverse events with potentially delayed onset, including rash, diarrhea, and colitis, cumulative exposure over the first 63 days (AUC_0\u0026ndash;63 days) was used, corresponding to the period during which drug concentrations were expected to approach steady state. Exposure\u0026ndash;safety analyses were stratified by QW and Q3W dosing regimens. Exploratory box plots were generated to visualize the distribution of exposure metrics across adverse event grades.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eBaseline and pharmacodynamic assessments\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003ePharmacodynamic assessments were conducted using whole blood, plasma, and tumor tissue samples, all of which were analyzed centrally. Whole blood and plasma samples were collected to evaluate changes in soluble biomarkers, cytokines, and immune-cell subsets using validated flow cytometry and immunoassay methods. Tumor tissue samples were analyzed by immunohistochemistry (IHC), immunofluorescence, and whole-transcriptome RNA sequencing to assess biomarkers of tumor inflammation and immune infiltration, including CD8⁺ tumor-infiltrating lymphocytes (TILs) and immune-related gene signatures.\u003c/p\u003e\n\u003cp\u003eOn-treatment tumor biopsies for the combination study were obtained on Cycle 2 Day 2 (Day 23). In the cibisatamab monotherapy study, on-treatment biopsies were collected at either Cycle 2 Day 1 (Day 21; Q3W dosing) or Cycle 7 Day 1 (Day 43; QW dosing), reflecting differences in treatment schedules.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eCytokines, soluble 4-1BB and CEA assessment\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eBlood samples were collected at baseline and at 1, 8, 15, 22, 29, 36, and 43 days following the first administration of FAP-4-1BBL. Peripheral inflammatory cytokines, soluble IL-2 receptor (sCD25), soluble 4-1BB, and serum CEA were quantified using validated ELISA-based platforms, including ProteinSimple ELLA cartridges for IFN-\u0026gamma;, IL-6, IL-2 receptor, and TNF-\u0026alpha;, Peprotech ELISA for soluble 4-1BB, and the Elecsys CEA assay (Cobas E170).\u003c/p\u003e\n\u003cp\u003eFor cytokine analyses, values below the lower limit of quantification (LLOQ) were imputed as 0.5 \u0026times; LLOQ with the addition of a small random noise term to facilitate model convergence, followed by log2 transformation (log2[value + 1]). Soluble CEA values were analyzed without imputation and were log10-transformed. Soluble 4-1BB values below LLOQ were handled similarly to cytokines and log10-transformed. All analyses were performed on changes from baseline.\u003c/p\u003e\n\u003cp\u003eRandom-effects models were used for longitudinal analyses, with participant specified as a random effect and visit timepoint as a fixed effect. Correction for multiple testing was performed using the false discovery rate (FDR), with statistical significance defined as FDR \u0026lt; 0.05.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eFlow cytometry\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003ePeripheral blood samples were collected in sodium heparin tubes and processed according to validated assay protocols. All staining and incubation steps were performed at room temperature in the dark. Data acquisition was performed using BD FACSCanto II instruments, with analysis conducted using FACSDiva software and assay-specific acquisition templates. Cell preparation was automated using a BD FACS Lyse Wash Assistant.\u003c/p\u003e\n\u003cp\u003eSurface staining was performed using antibody cocktails applied to whole blood samples, followed by red blood cell lysis, washing, and resuspension in phosphate-buffered saline. For intracellular staining, samples were additionally permeabilized using Perm Buffer II prior to incubation with intracellular antibody cocktails. A complete list of antibody reagents is provided in Supplementary Table 9.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eImmunohistochemistry\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eTumor tissue was collected at baseline and on Cycle 2 Day 2 (\u0026plusmn;2 days where feasible), preferentially from the same lesion. All samples were processed and paraffin-embedded according to standardized histopathology protocols. FFPE tissue sections were stained for hematoxylin and eosin, IHC, and immunofluorescence at Discovery Life Sciences (Kassel, Germany) or Roche Tissue Diagnostics (Tucson, Arizona). Only samples meeting predefined quality-control criteria for tumor content and tissue integrity were included. Details of staining assays are provided in Supplementary Table 10.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eCD3/CD8/Ki67/4-1BB/OX40 multiplex immunofluorescence\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eMultiplex immunofluorescence staining using five primary antibodies and tyramide signal amplification fluorophores was performed on a Ventana DISCOVERY ULTRA IHC/ISH platform, as previously described \u003csup\u003e54\u003c/sup\u003e. Whole-slide imaging was performed using a Zeiss Axioscan Z1 scanner. Quantitative image analysis was conducted using a custom-developed algorithm implemented in HALO (Indica Labs). Antibody and fluorophore details are provided in Supplementary Table 11.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eStatistics for immunohistochemistry and immunofluorescence analyses\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003ePharmacodynamic analyses were restricted to participants with matched baseline and on-treatment tumor samples. Paired changes were assessed using Wilcoxon matched-pairs signed-rank tests. Comparisons between pooled CR/PR participants and SD/PD participants were conducted using unpaired Mann\u0026ndash;Whitney tests. Additional comparisons were performed between participants with progressive disease and all remaining participants. All analyses were performed using GraphPad Prism version 10.0.0.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eStatistics for baseline associations with response\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eAssociations between baseline biomarkers and clinical response were evaluated using unpaired Mann\u0026ndash;Whitney tests for continuous variables. Fisher\u0026rsquo;s exact test was used to assess associations between liver metastases, consensus molecular subtype (CMS), and objective response. Progression-free survival (PFS) and overall survival (OS) were estimated using Kaplan\u0026ndash;Meier methods with 95% confidence intervals, and group comparisons were performed using log-rank tests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGene expression analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eRNA sequencing and processing\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eRNA was isolated from FFPE tumor tissue using the Qiagen AllPrep DNA/RNA FFPE kit. Library preparation was performed using the Illumina TruSeq RNA Exome kit. Base calling was conducted using bcl2fastq2, and sequence quality was assessed using FastQC \u003csup\u003e55\u003c/sup\u003e. Paired-end reads were aligned to the human genome (hg38) using STAR. Quality control was performed using MultiQC \u003csup\u003e54\u003c/sup\u003e. Gene-level read counts were generated using featureCounts \u003csup\u003e56\u003c/sup\u003e and normalized to transcripts per million (TPM).\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eDifferential gene expression and signature analyses\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eDifferential gene expression analyses were conducted using the limma package with log-transformed TPM values. Single-sample gene signature scores were calculated as the mean Z score of constituent genes. For signatures with up- and downregulated components, Z scores for downregulated genes were sign-inverted. Group comparisons were performed using limma.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eAdjustment of p-values\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eRaw p-values were adjusted for multiple testing using the Benjamini\u0026ndash;Hochberg procedure\u003csup\u003e57\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003ectDNA analysis\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eCirculating tumor DNA was analyzed using a personalized, tumor-informed, 16-plex PCR next-generation sequencing assay (Signatera RUO, Natera) \u003csup\u003e58\u003c/sup\u003e. Patient-specific somatic variants were identified from whole-exome sequencing data and used to design multiplex PCR assays. Plasma samples with \u0026ge;2 detected variants were classified as ctDNA-positive, and ctDNA levels were reported as mean tumor molecules per milliliter (MTM/ml).\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eWhole-exome sequencing\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eWhole-exome sequencing was performed on DNA extracted from macrodissected FFPE tumor tissue and matched germline DNA from peripheral blood, as previously described \u003csup\u003e58\u003c/sup\u003e. Reads were aligned to hg38 using BWA, duplicates were removed, and base quality score recalibration was performed using GATK \u003csup\u003e59\u003c/sup\u003e. Somatic variants were identified using Mutect2 and annotated with VEP \u003csup\u003e60\u003c/sup\u003e. Tumor mutational burden was calculated after standard filtering and normalized per megabase \u003csup\u003e61,62\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cu\u003ectDNA data analysis and statistics\u003c/u\u003e\u003c/p\u003e\n\u003cp\u003eKaplan\u0026ndash;Meier analyses for PFS and OS were performed using survfit2 and visualized using ggsurvfit. Participants were stratified into high and low ctDNA groups based on median baseline values. Landmark analyses at Cycle 3 Day 1 were applied to minimize immortal-time bias, with follow-up re-anchored at the landmark timepoint.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements and Disclosures\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank the patients for their participation in the study; the study nurses and site coordinators for their professionalism and dedication in supporting the conduct of the study; Christina Claus, Christian Klein, Vaios Karanikas (Roche Innovation Center Zurich) and Marie-H\u0026eacute;l\u0026egrave;ne Wasmer (Roche Innovation Center Basel) for continuous support and valuable discussions; Alfred Geiger (pRED Basel) for ensuring drug availability; Marleen Wilde, Nadine Kumpesa, and Marine Cordonnier (pRED Basel) for their contributions to study set-up; and Chiahuey Oii (pRED Basel) for providing the initial RNA-seq analysis. Sietzke Heyn for editorial assistance with the preparation of this article; and finally the numerous anonymous persons at the research sites as well as at Roche that have also been essential in supporting this project.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMedical writing and editorial assistance were provided by Dr. Raja Choudhury in accordance with Good Publication Practice (GPP3) guidelines and was funded by F. Hoffmann-La Roche Ltd.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eI.M, E.C.A, C.Q, M.D, I.B, S.B, F.T, M.C.R, M.M.G, V.W., T.W.K, V.M, were clinical investigators on the study.\u0026nbsp;A.B, E.G, C.H, T.T, H.H, C.M, P.U and I.M planned, designed and refined the study. A.B. lead the Roche study team. T.T., E.G and A.B wrote the initial draft of the paper, and all authors reviewed and edited the paper. C.H, I.D and T.T conducted biomarker analyses. C.M conducted pharmacokinetics analysis. L.C. developed the clinical analysis plan and analyzed data. H.H conducted the safety analysis. E.G., E.C., and O.K. provided medical support and oversight. T.S. had the operations oversight\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis work was supported by F. Hoffmann-La Roche Ltd, which significantly contributed to the study design in collaboration with study investigators, contributed to the collection of data, significantly contributed to the analysis and interpretation of the data in collaboration with study investigators, contributed to the writing of the report with study collaborators, and agreed with the study investigators in the decision to submit the paper for publication. No grant number is applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSeveral authors are employees of F. Hoffmann-La Roche Ltd and/or Genentech, Inc., and may hold Roche stock or stock options as part of their employment compensation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIosune Baraibar reports accommodation and travel expenses from Amgen, Merck, Sanofi, and Servier; and personal speaker honoraria from Astra Zeneca and Amgen.\u003c/p\u003e\n\u003cp\u003eVictor Morena declared financial and non-financial interests with Abbvie, Achilles, Adaptimmune, Adc Therapeutics, Ascendis Pharma, Astrazeneca, Bayer, Beigene, \u0026nbsp;Bicycle Tx, Bioinvent, Biomea Fusion, Biontech, Bms,, Boehringer, C4 Therapeutics, Calico Life Sciences Llc, Celgene, Constellation, Crescendo Biologics, Cullinan, Daiichi Sankyo, Debiopharm, Dragonfly, Enliven Therapeutics, Epizyme, Exelixis, Famewave, F-Star Beta Limited, Genentech, Genmab, Gilead, Grey Wolf Therapeutics, Gsk, Hexal Ag \u0026amp; Sandoz, Hifibio, Hookipa Biotech, Hutchmed, Igm Biosciences, Imcheck Therapeutics, Immunocore, Immutep, Incyte Iomx Therapeutics, Iovance, Italfarmaco, Iteos, Janssen, Light Chain Bioscience, Lilly, Loxo Oncology, Merck, Merus, Miltenyi Biomedicine, Monta Biosciences, Msd, Mythic Therapeutics, Ningbo Newbay, Novartis, Oxford Biotherapeutics, Pfizer, Pharmamar, Pmv Pharma, Prelude Therapeutics Inc, Pyxis Oncology, Regeneron, Relay Terapeutics, Repare Therapeutics, Revolution, Roche, Schr\u0026ouml;dinger, Scorpion Therapeutics, Seagen, Shattucks, Synthorx, Takeda, Tango Therapeutics, Tesaro, Totus Medicines, Turning Point Therapeutics, Vividion Therapeutics.\u003c/p\u003e\n\u003cp\u003eFiona Thistlethwaite declared institutional funding from\u0026nbsp;Achilles Ltd, Adaptimmune, Amgen, Biontech, BMS, Byondis, Chugai, Corbus, Crescendo Biologics, Grey Wolf Therapeutics, GSK, Immunocore, Iovance, Janssen, Kymab Ltd/Sanofi, Leucid, Moderna, Novalgen, Nucana, Oxford Vacmedix, Roche, RS Oncology LLC, Seagen, Takeda, T-Knife Therapeutics, UCB, Zelluna, Zymeworks. Advisory board/consultancy/honoraria: AstraZeneca, Grey Wolf Therapeutics, OncoBayes, T-Knife Therapeutics, Waypoint DSMB: Immatics\u003c/p\u003e\n\u003cp\u003eIgnacio Melero declared the following interests: Grants: Roche, AstraZenca, Bristol Myers, Genmab. Consultancy: Roche, Genmab, Regeneron, Pioneer medicines, Bright peak, Pharmamar, Curon, Mestag, Light chain, \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eEmiliano Calvo reports the following interests: ownership interests in Oncoart Associated; honoraria from HM Hospitales Group; had consulting/advisory role at Nanobiotix, Janssen-Cilag, Roche/Genentech, TargImmune Therapeutics, Servier, Bristol-Myers Squibb, Amunix, Adcendo, Anaveon, AstraZeneca/MedImmune, Chugai Pharma, MonTa, MSD Oncology, Nouscom, Novartis, OncoDNA, T-Knife, Elevation Oncology, PharmaMar, Ellipses Pharma, Syneos Health, Genmab and Diaccurate; is president and founder of Foundation INTHEOS.\u003c/p\u003e\n\u003cp\u003eTae Won Kim declared institutional funding from Genentech and Inocras.\u003c/p\u003e\n\u003cp\u003eSeung Hoon Beom declared the following interests: Researcher/Grant/Contract (clinical trial PI): Roche, MSD, Bayer, Boryung, Takeda, Jeil Pharmaceutical, Merck, Janssen, Pfizer, Amgen, Revolution Medicines, IGM Biosciences, STcube\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAlouani, E., \u003cem\u003eet al.\u003c/em\u003e Efficacy of immunotherapy in mismatch repair-deficient advanced colorectal cancer in routine clinical practice. An AGEO study. \u003cem\u003eESMO Open\u003c/em\u003e 8, 101574 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, E. \u0026amp; Zhou, W. Immunotherapy in microsatellite-stable colorectal cancer: Strategies to overcome resistance. \u003cem\u003eCrit Rev Oncol Hematol\u003c/em\u003e 212, 104775 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCai, L., Chen, A. \u0026amp; Tang, D. A new strategy for immunotherapy of microsatellite-stable (MSS)-type advanced colorectal cancer: Multi-pathway combination therapy with PD-1/PD-L1 inhibitors. \u003cem\u003eImmunology\u003c/em\u003e 173, 209\u0026ndash;226 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorano, F., \u003cem\u003eet al.\u003c/em\u003e Temozolomide Followed by Combination With Low-Dose Ipilimumab and Nivolumab in Patients With Microsatellite-Stable, O(6)-Methylguanine-DNA Methyltransferase-Silenced Metastatic Colorectal Cancer: The MAYA Trial. \u003cem\u003eJ Clin Oncol\u003c/em\u003e 40, 1562\u0026ndash;1573 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHutchings, M., \u003cem\u003eet al.\u003c/em\u003e Glofitamab, a Novel, Bivalent CD20-Targeting T-Cell-Engaging Bispecific Antibody, Induces Durable Complete Remissions in Relapsed or Refractory B-Cell Lymphoma: A Phase I Trial. \u003cem\u003eJ Clin Oncol\u003c/em\u003e 39, 1959\u0026ndash;1970 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbramson, J.S., \u003cem\u003eet al.\u003c/em\u003e Glofitamab plus gemcitabine and oxaliplatin (GemOx) versus rituximab-GemOx for relapsed or refractory diffuse large B-cell lymphoma (STARGLO): a global phase 3, randomised, open-label trial. \u003cem\u003eLancet\u003c/em\u003e 404, 1940\u0026ndash;1954 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePaz-Ares, L., \u003cem\u003eet al.\u003c/em\u003e Tarlatamab, a First-in-Class DLL3-Targeted Bispecific T-Cell Engager, in Recurrent Small-Cell Lung Cancer: An Open-Label, Phase I Study. \u003cem\u003eJ Clin Oncol\u003c/em\u003e 41, 2893\u0026ndash;2903 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMountzios, G., \u003cem\u003eet al.\u003c/em\u003e Tarlatamab in Small-Cell Lung Cancer after Platinum-Based Chemotherapy. \u003cem\u003eN Engl J Med\u003c/em\u003e 393, 349\u0026ndash;361 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, L. \u0026amp; Flies, D.B. Molecular mechanisms of T cell co-stimulation and co-inhibition. \u003cem\u003eNat Rev Immunol\u003c/em\u003e 13, 227\u0026ndash;242 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChester, C., Sanmamed, M.F., Wang, J. \u0026amp; Melero, I. Immunotherapy targeting 4-1BB: mechanistic rationale, clinical results, and future strategies. \u003cem\u003eBlood\u003c/em\u003e 131, 49\u0026ndash;57 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelero, I., \u003cem\u003eet al.\u003c/em\u003e CD137 (4-1BB)-Based Cancer Immunotherapy on Its 25th Anniversary. \u003cem\u003eCancer Discov\u003c/em\u003e 13, 552\u0026ndash;569 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelero, I., \u003cem\u003eet al.\u003c/em\u003e Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors. \u003cem\u003eNat Med\u003c/em\u003e 3, 682\u0026ndash;685 (1997).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTiernan, J.P., \u003cem\u003eet al.\u003c/em\u003e Carcinoembryonic antigen is the preferred biomarker for in vivo colorectal cancer targeting. \u003cem\u003eBr J Cancer\u003c/em\u003e 108, 662\u0026ndash;667 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBacac, M., \u003cem\u003eet al.\u003c/em\u003e A Novel Carcinoembryonic Antigen T-Cell Bispecific Antibody (CEA TCB) for the Treatment of Solid Tumors. \u003cem\u003eClin Cancer Res\u003c/em\u003e 22, 3286\u0026ndash;3297 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLehmann, S., \u003cem\u003eet al.\u003c/em\u003e In Vivo Fluorescence Imaging of the Activity of CEA TCB, a Novel T-Cell Bispecific Antibody, Reveals Highly Specific Tumor Targeting and Fast Induction of T-Cell-Mediated Tumor Killing. \u003cem\u003eClin Cancer Res\u003c/em\u003e 22, 4417\u0026ndash;4427 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSegal, N.H., \u003cem\u003eet al.\u003c/em\u003e CEA-CD3 bispecific antibody cibisatamab with or without atezolizumab in patients with CEA-positive solid tumours: results of two multi-institutional Phase 1 trials. \u003cem\u003eNat Commun\u003c/em\u003e 15, 4091 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeijeira, A., \u003cem\u003eet al.\u003c/em\u003e Three-dimensional colon cancer organoids model the response to CEA-CD3 T-cell engagers. \u003cem\u003eTheranostics\u003c/em\u003e 12, 1373\u0026ndash;1387 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClaus, C., \u003cem\u003eet al.\u003c/em\u003e Tumor-targeted 4-1BB agonists for combination with T cell bispecific antibodies as off-the-shelf therapy. \u003cem\u003eSci Transl Med\u003c/em\u003e 11(2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeters, S., \u003cem\u003eet al.\u003c/em\u003e Obinutuzumab Pretreatment as a Novel Approach to Mitigate Formation of Anti-Drug Antibodies Against Cergutuzumab Amunaleukin in Patients with Solid Tumors. \u003cem\u003eClin Cancer Res\u003c/em\u003e 30, 1630\u0026ndash;1641 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanchez, J., \u003cem\u003eet al.\u003c/em\u003e Combining mathematical modeling, in vitro data and clinical target expression to support bispecific antibody binding affinity selection: a case example with FAP-4-1BBL. \u003cem\u003eFront Pharmacol\u003c/em\u003e 15, 1472662 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuinney, J., \u003cem\u003eet al.\u003c/em\u003e The consensus molecular subtypes of colorectal cancer. \u003cem\u003eNat Med\u003c/em\u003e 21, 1350\u0026ndash;1356 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCiardiello, F., \u003cem\u003eet al.\u003c/em\u003e Clinical management of metastatic colorectal cancer in the era of precision medicine. \u003cem\u003eCA Cancer J Clin\u003c/em\u003e 72, 372\u0026ndash;401 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHegde, P.S., Karanikas, V. \u0026amp; Evers, S. The Where, the When, and the How of Immune Monitoring for Cancer Immunotherapies in the Era of Checkpoint Inhibition. \u003cem\u003eClin Cancer Res\u003c/em\u003e 22, 1865\u0026ndash;1874 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMisiakos, E.P., Karidis, N.P. \u0026amp; Kouraklis, G. Current treatment for colorectal liver metastases. \u003cem\u003eWorld J Gastroenterol\u003c/em\u003e 17, 4067\u0026ndash;4075 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, J., \u003cem\u003eet al.\u003c/em\u003e Liver metastasis restrains immunotherapy efficacy via macrophage-mediated T cell elimination. \u003cem\u003eNat Med\u003c/em\u003e 27, 152\u0026ndash;164 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGlez-Vaz, J., \u003cem\u003eet al.\u003c/em\u003e Soluble CD137 as a dynamic biomarker to monitor agonist CD137 immunotherapies. \u003cem\u003eJ Immunother Cancer\u003c/em\u003e 10(2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGeorganaki, M., \u003cem\u003eet al.\u003c/em\u003e Tumor endothelial cell up-regulation of IDO1 is an immunosuppressive feed-back mechanism that reduces the response to CD40-stimulating immunotherapy. \u003cem\u003eOncoimmunology\u003c/em\u003e 9, 1730538 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJenkins, E., Whitehead, T., Fellermeyer, M., Davis, S.J. \u0026amp; Sharma, S. The current state and future of T-cell exhaustion research. \u003cem\u003eOxf Open Immunol\u003c/em\u003e 4, iqad006 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorny\u0026aacute;k, L., \u003cem\u003eet al.\u003c/em\u003e The Role of Indoleamine-2,3-Dioxygenase in Cancer Development, Diagnostics, and Therapy. \u003cem\u003eFront Immunol\u003c/em\u003e 9, 151 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMunn, D.H. \u0026amp; Mellor, A.L. Indoleamine 2,3-dioxygenase and tumor-induced tolerance. \u003cem\u003eJ Clin Invest\u003c/em\u003e 117, 1147\u0026ndash;1154 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBretscher, P.A. The history of the two-signal model of lymphocyte activation: A personal perspective. \u003cem\u003eScand J Immunol\u003c/em\u003e 89, e12762 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEtxeberria, I., \u003cem\u003eet al.\u003c/em\u003e Engineering bionic T cells: signal 1, signal 2, signal 3, reprogramming and the removal of inhibitory mechanisms. \u003cem\u003eCell Mol Immunol\u003c/em\u003e 17, 576\u0026ndash;586 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWatts, T.H., Yeung, K.K.M., Yu, T., Lee, S. \u0026amp; Eshraghisamani, R. TNF/TNFR Superfamily Members in Costimulation of T Cell Responses-Revisited. \u003cem\u003eAnnu Rev Immunol\u003c/em\u003e 43, 113\u0026ndash;142 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDziadek, S., \u003cem\u003eet al.\u003c/em\u003e Comprehensive analysis of fibroblast activation protein expression across 23 tumor indications: insights for biomarker development in cancer immunotherapies. \u003cem\u003eFront Immunol\u003c/em\u003e 15, 1352615 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKraxner, A., \u003cem\u003eet al.\u003c/em\u003e Investigating the complex interplay between fibroblast activation protein α-positive cancer associated fibroblasts and the tumor microenvironment in the context of cancer immunotherapy. \u003cem\u003eFront Immunol\u003c/em\u003e 15, 1352632 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCabrita, R., \u003cem\u003eet al.\u003c/em\u003e Tertiary lymphoid structures improve immunotherapy and survival in melanoma. \u003cem\u003eNature\u003c/em\u003e 577, 561\u0026ndash;565 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConejo-Garcia, J.R., Biswas, S., Chaurio, R. \u0026amp; Rodriguez, P.C. Neglected no more: B cell-mediated anti-tumor immunity. \u003cem\u003eSemin Immunol\u003c/em\u003e 65, 101707 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFridman, W.H., \u003cem\u003eet al.\u003c/em\u003e Tertiary lymphoid structures and B cells: An intratumoral immunity cycle. \u003cem\u003eImmunity\u003c/em\u003e 56, 2254\u0026ndash;2269 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Visser, K.E., Korets, L.V. \u0026amp; Coussens, L.M. De novo carcinogenesis promoted by chronic inflammation is B lymphocyte dependent. \u003cem\u003eCancer Cell\u003c/em\u003e 7, 411\u0026ndash;423 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMenk, A.V., \u003cem\u003eet al.\u003c/em\u003e 4-1BB costimulation induces T cell mitochondrial function and biogenesis enabling cancer immunotherapeutic responses. \u003cem\u003eJ Exp Med\u003c/em\u003e 215, 1091\u0026ndash;1100 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTr\u0026uuml;b, M., \u003cem\u003eet al.\u003c/em\u003e Fibroblast activation protein-targeted-4-1BB ligand agonist amplifies effector functions of intratumoral T cells in human cancer. \u003cem\u003eJ Immunother Cancer\u003c/em\u003e 8(2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSegal, N.H., \u003cem\u003eet al.\u003c/em\u003e Results from an Integrated Safety Analysis of Urelumab, an Agonist Anti-CD137 Monoclonal Antibody. \u003cem\u003eClin Cancer Res\u003c/em\u003e 23, 1929\u0026ndash;1936 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelero, I., \u003cem\u003eet al.\u003c/em\u003e A first-in-human study of the fibroblast activation protein-targeted, 4-1BB agonist RO7122290 in patients with advanced solid tumors. \u003cem\u003eSci Transl Med\u003c/em\u003e 15, eabp9229 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelero, I., \u003cem\u003eet al.\u003c/em\u003e 151P Enhanced pharmacodynamic effects upon combination of cibisatamab and FAP-4-1BBL in 3L+ mMSS CRC patients. \u003cem\u003eAnnals of Oncology\u003c/em\u003e 35, S275-S276 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee, H.H. \u0026amp; Al-Ogaili, Z. Fibroblast activation protein and the tumour microenvironment: challenges and therapeutic opportunities. \u003cem\u003eOncol Rev\u003c/em\u003e 19, 1617487 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelero, I., \u003cem\u003eet al.\u003c/em\u003e Amplification of tumor immunity by gene transfer of the co-stimulatory 4-1BB ligand: synergy with the CD28 co-stimulatory pathway. \u003cem\u003eEur J Immunol\u003c/em\u003e 28, 1116\u0026ndash;1121 (1998).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYe, Z., \u003cem\u003eet al.\u003c/em\u003e Gene therapy for cancer using single-chain Fv fragments specific for 4-1BB. \u003cem\u003eNat Med\u003c/em\u003e 8, 343\u0026ndash;348 (2002).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBr\u0026ouml;ske, A.E., \u003cem\u003eet al.\u003c/em\u003e Pharmacodynamics and molecular correlates of response to glofitamab in relapsed/refractory non-Hodgkin lymphoma. \u003cem\u003eBlood Adv\u003c/em\u003e 6, 1025\u0026ndash;1037 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO'Brien, N., \u003cem\u003eet al.\u003c/em\u003e T cell margination: investigating the detour of T cells following forimtamig treatment in humanized mice. \u003cem\u003eMAbs\u003c/em\u003e 17, 2440578 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStrimpakos, A.S., \u003cem\u003eet al.\u003c/em\u003e The impact of carcinoembryonic antigen flare in patients with advanced colorectal cancer receiving first-line chemotherapy. \u003cem\u003eAnn Oncol\u003c/em\u003e 21, 1013\u0026ndash;1019 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCallesen, L.B., \u003cem\u003eet al.\u003c/em\u003e Circulating tumour DNA and its clinical utility in predicting treatment response or survival in patients with metastatic colorectal cancer: a systematic review and meta-analysis. \u003cem\u003eBr J Cancer\u003c/em\u003e 127, 500\u0026ndash;513 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTie, J., \u003cem\u003eet al.\u003c/em\u003e Circulating Tumor DNA Analyses as Markers of Recurrence Risk and Benefit of Adjuvant Therapy for Stage III Colon Cancer. \u003cem\u003eJAMA Oncol\u003c/em\u003e 5, 1710\u0026ndash;1717 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSharma, P., Hu-Lieskovan, S., Wargo, J.A. \u0026amp; Ribas, A. Primary, Adaptive, and Acquired Resistance to Cancer Immunotherapy. \u003cem\u003eCell\u003c/em\u003e 168, 707\u0026ndash;723 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEwels, P., Magnusson, M., Lundin, S. \u0026amp; K\u0026auml;ller, M. MultiQC: summarize analysis results for multiple tools and samples in a single report. \u003cem\u003eBioinformatics\u003c/em\u003e 32, 3047\u0026ndash;3048 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndrews, S., Gilley, J. \u0026amp; Coleman, M.P. Difference Tracker: ImageJ plugins for fully automated analysis of multiple axonal transport parameters. \u003cem\u003eJ Neurosci Methods\u003c/em\u003e 193, 281\u0026ndash;287 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiao, Y., Smyth, G.K. \u0026amp; Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. \u003cem\u003eBioinformatics\u003c/em\u003e 30, 923\u0026ndash;930 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBenjamini, Y. \u0026amp; Hochberg, Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. \u003cem\u003eJournal of the Royal Statistical Society: Series B (Methodological)\u003c/em\u003e 57, 289\u0026ndash;300 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eReinert, T., \u003cem\u003eet al.\u003c/em\u003e Analysis of Plasma Cell-Free DNA by Ultradeep Sequencing in Patients With Stages I to III Colorectal Cancer. \u003cem\u003eJAMA Oncol\u003c/em\u003e 5, 1124\u0026ndash;1131 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAuwera, G.v.d. \u0026amp; O'Connor, B.D. \u003cem\u003eGenomics in the cloud: using Docker, GATK, and WDL in Terra\u003c/em\u003e, (O'Reilly Media, Sebastopol, CA, 2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcLaren, W., \u003cem\u003eet al.\u003c/em\u003e The Ensembl Variant Effect Predictor. \u003cem\u003eGenome Biol\u003c/em\u003e 17, 122 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarczewski, K.J., \u003cem\u003eet al.\u003c/em\u003e The mutational constraint spectrum quantified from variation in 141,456 humans. \u003cem\u003eNature\u003c/em\u003e 581, 434\u0026ndash;443 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmemiya, H.M., Kundaje, A. \u0026amp; Boyle, A.P. The ENCODE Blacklist: Identification of Problematic Regions of the Genome. \u003cem\u003eSci Rep\u003c/em\u003e 9, 9354 (2019).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8628656/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8628656/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCibisatamab is a carcinoembryonic antigen (CEA)\u0026ndash;directed T-cell engaging bispecific antibody that has shown evidence of antitumor activity as a single agent in patients with microsatellite-stable (MSS) metastatic colorectal cancer (mCRC) (ClinicalTrials.gov identifier: NCT02324257). Effective T-cell activation requires coordinated signaling through the TCR\u0026ndash;CD3 complex and costimulatory pathways, including 4-1BB, which is induced following antigen recognition. We evaluated the combination of cibisatamab with escalating doses of FAP-4-1BBL, a fibroblast activation protein (FAP)\u0026ndash;targeted 4-1BB agonist, in an open-label phase 1b study (BP42675; ClinicalTrials.gov identifier: NCT04826003) in patients with MSS mCRC who had progressed after at least two prior lines of therapy.\u003c/p\u003e \u003cp\u003eThe combination demonstrated a manageable safety profile and induced robust immune activation, characterized by sustained increases in circulating interferon-γ (IFNγ), soluble CD25 (sCD25), and proliferating/ activated CD8⁺ T-cell subsets, together with expansion of memory compartments. Compared with cibisatamab monotherapy, the combination resulted in greater and more durable induction of IFNγ and sCD25 without any exacerbation of cytokine-release\u0026ndash;related toxicity. Treatment was associated with numerically higher overall response and disease control rates than previously reported for cibisatamab alone, with an overall response rate of 17.6% (9/51) and a disease control rate of 50.9% (26/51) across dose levels and schedules. Reductions in serum CEA, early decreases in circulating tumor DNA, and increased soluble CD137 were associated with clinical benefit. Paired tumor biopsies showed greater increases of intratumoral CD8⁺ and CD8⁺Ki67⁺ T-cell infiltration relative to cibisatamab monotherapy.\u003c/p\u003e \u003cp\u003e Together, these findings support tumor-localized 4-1BB costimulation as a strategy to enhance the biological and clinical activity of T-cell engagers in MSS mCRC and other non-inflamed solid tumors.\u003c/p\u003e","manuscriptTitle":"Tumor-targeted 4-1BB costimulation enhances immune activation and clinical activity of a CEA-directed T-cell engager in microsatellite-stable colorectal cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-09 12:57:12","doi":"10.21203/rs.3.rs-8628656/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-medicine","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nm","sideBox":"Learn more about [Nature Medicine](http://www.nature.com/nm/)","snPcode":"","submissionUrl":"","title":"Nature Medicine","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"d195f064-2044-4488-8e2c-df0fb425e720","owner":[],"postedDate":"February 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":61440145,"name":"Health sciences/Medical research"},{"id":61440146,"name":"Health sciences/Medical research/Outcomes research"}],"tags":[],"updatedAt":"2026-04-21T07:06:26+00:00","versionOfRecord":{"articleIdentity":"rs-8628656","link":"https://doi.org/10.1038/s41591-026-04380-z","journal":{"identity":"nature-medicine","isVorOnly":false,"title":"Nature Medicine"},"publishedOn":"2026-04-20 04:00:00","publishedOnDateReadable":"April 20th, 2026"},"versionCreatedAt":"2026-02-09 12:57:12","video":"","vorDoi":"10.1038/s41591-026-04380-z","vorDoiUrl":"https://doi.org/10.1038/s41591-026-04380-z","workflowStages":[]},"version":"v1","identity":"rs-8628656","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8628656","identity":"rs-8628656","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}