Search What Studies Actually Found

Designing a Phase I/II Multiantigen CAR T Trial Protocol for Solid Tumors

Below is an evidence based, experimentally actionable clinical trial protocol template focused on multiantigen CAR T strategies to reduce antigen escape in solid tumors, with manufacturing, safety monitoring, correlative science, and statistical plans. All claims are supported by recent literature. This is a translational protocol design and not clinical advice.

Executive summary

\n
• Rationale: antigen heterogeneity and antigen loss limit single antigen CAR T efficacy in solid tumors; multiantigen CAR designs (bicistronic, tandem, logic gated) reduce escape and can improve specificity and function CAR
\n
• Target selection example: prioritize pairs with complementary expression and nonoverlapping normal tissue expression (eg B7-H3 plus GD2 or HER2 plus MUC1) and confirm by IHC/RNA data in the intended tumor type because spatial localization matters for efficacy and toxicity B7-H3
\n
• Manufacturing: consider bicistronic or tandem CAR T architectures with costimulatory domains chosen to tune expansion and toxicity (CD28 vs 4-1BB), and include modular safety features (suicide switch or OFF adaptor) and optional armored payloads (eg IL-18/IL-15) if justified by preclinical safety data Multiplex
\n

1 Trial objectives

\n
1. Primary Phase I: determine recommended phase II dose (RP2D) and dose limiting toxicities (DLTs) for multiantigen CAR T product.
\n
2. Secondary Phase I/II: assess objective response rate (ORR) per RECIST v1.1 (or disease specific criteria), duration of response (DoR), progression-free survival (PFS), overall survival (OS), and safety (CRS, ICANS, on-target off-tumor toxicity).
\n
3. Exploratory: evaluate antigen expression dynamics, CAR T persistence, tumor immune microenvironment remodeling, and biomarker correlates predictive of response (IHC, RNA, single cell where feasible).
\n

2 Key scientific hypothesis

Simultaneous targeting of two validated tumor antigens using a bicistronic or tandem CAR T architecture will reduce antigen escape and increase tumor control with an acceptable safety profile compared with historic single antigen CAR T results in the same indication, while modular control elements will limit off-tumor toxicity and permit dose escalation

3 Indication and target selection strategy

Protocol applies to patients with advanced/recurrent solid tumors of a single histology cohort (example: recurrent glioblastoma or metastatic pancreatic adenocarcinoma). Target pair selection criteria:

\n
• High prevalence of both antigens in tumor cohort by IHC and/or RNA (local validation required) and low expression in vital normal tissues
\n
• Complementary spatial expression to minimize escape (eg membrane targets accessible to CAR binding) and minimal nuclear/stromal localization that may evade CAR recognition B7-H3
\n
• Preclinical evidence of additive or synergistic cytotoxicity for the antigen pair and low cross-reactivity with normal tissues (literature and in vitro testing)
\n

4 CAR design options (choose per program)

5 Trial design overview

6 Eligibility (example)

Inclusion

\n
• Age 18 75 years; ECOG 0 2
\n
• Histologic confirmation of selected tumor type with at least one measurable lesion per RECIST v1.1
\n
• Central confirmation of expression of both target antigens by IHC above prespecified thresholds
\n
• Adequate organ function and ability to undergo leukapheresis
\n

Exclusion (high level)

\n
• Active uncontrolled infection, HIV or active hepatitis B/C without suppression, uncontrolled autoimmune disease that may be exacerbated by CAR T
\n
• Significant cardiac or pulmonary disease that increases risk for CRS/ICANS
\n
• Pregnancy or breastfeeding
\n
• Prior allogeneic HSCT within 6 months or active GVHD
\n

7 Manufacturing and release testing

\n
1. Apheresis per institutional SOPs; consider individualized apheresis volume algorithm to optimize collection and minimize machine time as per recent decentralized CAR T work Individualized
\n
2. Vector choice: clinical grade lentiviral for stable expression, or nonviral (eg transposon) if validated; include manufacturing limits for vector copy number, sterility, mycoplasma, replication competent virus testing, endotoxin, and identity (flow cytometry) prior to release.
\n
3. Product testing: transduction efficiency, viability, phenotype (CD4/CD8 ratio, memory phenotype), replication competent virus, absence of contamination, potency assay (in vitro killing or cytokine release against antigen positive targets). Document presence/function of safety switch and any cytokine payload expression if used.
\n

8 Dosing and escalation

Start low and escalate: a conservative 3+3 or Bayesian model with dose levels expressed as total CAR T cell dose per body weight (eg 1e6, 3e6, 1e7 CAR T cells/kg or fixed doses depending on prior data). Consider intra-patient dose escalation only in highly controlled settings. Include stopping rules for prespecified severe toxicity rates.

9 Safety management and monitoring

\n
• Inpatient observation for at least 7 days post-infusion with standardized CRS/ICANS grading per ASTCT and protocols for tocilizumab and steroids
\n
• Neurologic baseline assessment and daily monitoring with EEG and neuro checks for ICANS risk if locoregional CNS delivery
\n
• Infection prophylaxis and monitoring; specialized plan for on-target off-tumor organ injury detection (eg serial organ-specific labs and imaging)
\n
• Early use of suicide switch activation criteria in case of life-threatening off-tumor toxicity
\n

10 Correlative sciences and biomarkers

Implement a prespecified correlative science program to test the mechanistic hypotheses about multiantigen targeting, persistence, and microenvironment remodeling:

\n
1. Tumor biopsies pre-treatment and on-treatment (where feasible) for IHC of both antigens, immune infiltrate (CD8, CD4, Tregs, macrophage markers), spatial profiling, and RNAseq. B7-H3 biology shows RNA and localization matter; include both protein and RNA assays B7-H3
\n
2. Peripheral blood monitoring: serial flow cytometry for CAR positivity, phenotype, exhaustion markers; cell free DNA for tumor dynamics; serum cytokines (IL-6, IFN-gamma, IL-1, IL-18 if relevant).
\n
3. Single cell or bulk transcriptomics of infused cells and tumor (datasets E-MTAB-15147 and E-MTAB-12778 are relevant resources for CAR T transcriptional signatures and CCR7 engineering comparisons; include for comparative analyses) E\n
4. Functional assays: ex vivo cytotoxicity, antigen downregulation assays, and serum neutralization where relevant.
\n

11 Statistical considerations and sample size

Phase I: classic 3+3 or model-based escalation until MTD/RP2D determined; expected 12 24 patients depending on DLTs. Phase II expansion: for single arm efficacy, design to detect an ORR improvement from historical rate p0 to target p1 with one stage or Simon two-stage design. Example: if historical ORR 10% and target 30%, with alpha 0.05 and power 80% sample size ~29 evaluable patients (Simon two stage parameters to be calculated specifically for chosen indication).

12 Ethical and regulatory points

\n
• Pre-IND meeting with regulatory authority to agree manufacturing, release criteria, toxicity management and stopping rules.
\n
• Long term follow-up plan for insertional oncogenesis per gene therapy guidance and for 15 years if integrating vectors used.
\n
• Independent Data Safety Monitoring Board (DSMB) with stopping rules for unexpected severe toxicity.
\n

13 Preclinical data requirements before first-in-human

\n
• In vitro killing across antigen combinatorial expression levels and hinge/affinity optimization (hinge truncation can tune antigen sensitivity and improve safety margins; consider tuneable hinge length in design) Hinge
\n
• In vivo efficacy and toxicity in appropriate xenograft and immunocompetent models showing tumor control and acceptable organ toxicity, plus biodistribution to ensure lack of high-level normal tissue binding.
\n
• GLP toxicology including biodistribution and persistence (vector shedding if viral), at doses covering projected clinical exposures.
\n

14 Sample correlative timeline (recommended)

\n
1. Baseline tumor biopsy + blood pre-leukapheresis
\n
2. Product sample retained at cryobank (pre-infusion)
\n
3. Peripheral blood at day 0 pre-infusion, days 1, 3, 7, 14, 28, monthly to 6 months, q3 months to 2 years
\
4. Optional on-treatment tumor biopsy at week 4 6 where safe
\

15 Interpretational caveats and blindspots

\n
• Preclinical models imperfectly predict human toxicity and persistence; expect translational gaps and rigorously conservative dose escalation Translational
\n
• B7-H3 and other targets can have discordant RNA and protein localization; ensure protein IHC confirmation of membrane expression and not only transcript abundance RNA
\n
• Manufacturing variability and patient T-cell fitness affect product quality; collect manufacturing covariates and consider bridging therapy policies carefully.
\n

16 Implementation aids and datasets

Use public datasets to benchmark transcriptomic and single cell signatures and for comparator analyses. Examples included in correlative plan:

\n
• ArrayExpress E-MTAB-15147 for single cell CAR T transcriptional profiles in brain organoid GD2 models to compare infused vs persisting phenotypes E
\n
• ArrayExpress E-MTAB-12778 for bulk transcriptional changes introduced by CCR7 engineering as a model for evaluating chemokine receptor engineering effects on trafficking and phenotype E
\n

17 Suggested timelines

From IND-enabling studies to first-in-human: 12 24 months typical depending on GLP tox, vector manufacture scale up, and IRB approval. Early clinical readouts for safety within first 3 months and preliminary efficacy at 6 months.

18 Practical next steps to operationalize

\n
1. Assemble multidisciplinary team: cell therapy, pathology, neuro ICU if CNS indication, GMP manufacturing, DSMB, translational lab.
\n
2. Complete antigen validation by centralized IHC and RNA in target population and normal tissues.
\n
3. Perform optimized CAR design selection with hinge/linker tuning, costimulatory domain comparisons, and safety switch integration followed by GLP toxicology and biodistribution.
\n
4. Draft protocol with conservative dose escalation, robust CRS/ICANS management, and detailed correlative science consent and biobanking plans.
\n

19 Summary and confidence statement

This protocol template synthesizes contemporary, peer reviewed evidence supporting multiantigen CAR strategies to overcome antigen heterogeneity in solid tumors, draws on engineering advances to tune sensitivity and reduce exhaustion, and embeds multiple safety and correlative safeguards. Key supporting sources include a 2025 review of CAR T in solid tumors, a 2025 pan-cancer analysis of B7-H3 expression, CAR-NK design considerations relevant to safety, and preclinical engineering studies on bicistronic signaling and hinge tuning CARB7MultiplexDual