Research Projects

Our laboratory integrates target identification, mechanistic dissection, compound evaluation, and early drug-discovery validation within a single end-to-end pipeline. This page is structured in three parts: (1) the four research foundations that drive our discovery work; (2) the disease-focused programs in which those foundations are applied; and (3) partnership opportunities for industry and academic collaborators.

(Figure placeholder — replace with figure or remove before publication)

---

Our Research Foundations

Four integrated research foundations support our discovery pipeline, from target identification through mechanistic analysis and pharmacological evaluation.

---

Foundation 1 — Paralog Co-Inhibition Strategy and Paralog-Pair Dual-KO Library

1-to-2 Synthetic Lethality Strategy

Approximately one-third of human genes have functional paralogs — duplicated genes whose redundancy buffers the loss of any single member. Conventional synthetic-lethality screens focus on the two-factor relationship ("inhibit gene B, given that gene A is lost in cancer"); against paralog-buffered targets, single-gene inhibition often fails because the redundant paralog compensates.

We pioneered the paralog co-inhibition strategy — a 1-to-2 synthetic-lethality framework that simultaneously inhibits both members of a paralog pair that would otherwise buffer each other (Sasaki et al., Nat Commun 2024). Starting from a gene lost in cancer (factor 1), simultaneous inhibition of the paralog pair (factors 2 and 3) aims to exploit vulnerabilities specific to cancer cells carrying the relevant genetic alteration. This extends the conventional two-factor framework to a three-factor relationship and forms the conceptual basis for the higher-order frameworks described in Research Highlights Theme 3.

Assets underlying the strategy:

• Proprietary genome-scale paralog-pair dual-knockout vector library — not represented in public dependency databases. Enables recovery of paralog-buffered synthetic-lethal targets that single-gene screens systematically miss.
• CRISPR/Cas9 screening pipeline — in-house screen design and downstream analysis optimized for rare- and refractory-cancer cell lines.

Validated results (within published scope):

• CBP/p300 dual inhibition is synthetic-lethal in SMARCB1-deficient cancers (malignant rhabdoid tumor, epithelioid sarcoma) — Nat Commun 2024 (Sasaki et al.)
• Extension to cBAF-deficient cancers (SMARCA4-deficient lung cancer, synovial sarcoma) — Cancer Res Commun 2025 (Sasaki et al.)
• Research-use compound CP-C27 (dual CBP/p300 inhibitor) used for in vivo validation — Nat Commun 2024

→ Thematic deep dive: Research Highlights — Paralog Co-Inhibition

---

Foundation 2 — Rare- and Refractory-Cancer Cell-Line Panels

In-House Disease Models Covering Japan-Specific and Rare Cancer Contexts

Our in-house cell-line panels represent the genetic backgrounds central to our research programs:

• SMARCB1-deficient models (malignant rhabdoid tumor, epithelioid sarcoma)
• ARID1A-deficient models (ovarian clear cell carcinoma, diffuse-type gastric cancer)
• SMARCA4-deficient models (non-small cell lung cancer)
• KRAS-mutant models (lung cancer: G12C; pancreatic cancer: G12D, G12V)
• SMAD4-deficient models (diffuse gastric cancer, pancreatic cancer, esophageal cancer)
• KDM6A-deficient models (esophageal cancer)
• PBRM1-deficient models (renal cell carcinoma)

The panel includes patient-derived cell lines for gastric and pancreatic cancers, and comprehensive panels spanning lung cancer, sarcoma, esophageal cancer, and renal cancer — covering cancer types and molecular subtypes underrepresented in international public databases such as DepMap. These assets support disease-relevant validation where public dependency databases alone are insufficient.

→ Applied in all six disease programs and in DepMap-orthogonal validation workflows.

---

Foundation 3 — Data-Driven Target Discovery

DepMap Reanalysis Integrated with In-House CRISPR/Cas9 Screening

Public cancer dependency datasets contain informative signals that are obscured by pan-cancer averaging. We apply three analytical layers to extract rare-cancer-specific vulnerabilities:

1. Context-specific re-analysis — restricted to a single cancer lineage with molecular subgroup stratification (e.g., ARID1A-deficient vs. intact ovarian clear cell carcinoma)
2. Integration with in-house cell lines — orthogonal validation in cell lines absent from DepMap, including Japan-specific and rare cancer subtypes
3. Pathway-level interpretation — accounting for paralog-mediated redundancy that can suppress dependency signals in public data

In practice, we validate candidate dependencies from DepMap re-analysis by CRISPR/Cas9 screening in our in-house cell-line panels, comparing results against Chronos-based DepMap gene-effect scores.

Representative result:

• USP8 identified as a synthetic-lethal target in ARID1A-deficient ovarian clear cell carcinoma — NPJ Precis Oncol 2025 (Saito et al.)

→ Thematic deep dive: Research Highlights — Data-Driven Target Discovery

---

Foundation 4 — Mechanistic and Pharmacological Validation

Multi-Omics Mechanistic Analysis Connected to In Vitro and In Vivo Evaluation

Following target identification, we dissect the mechanism by which a vulnerability exists and validate its pharmacological tractability through:

• Multi-omics profiling — RNA-seq, ATAC-seq, CUT&RUN, and ChIP-seq for chromatin-level mechanistic dissection
• Metabolic analysis — metabolomics, glutathione metabolism, and pyrimidine-metabolism pathway analysis
• Pharmacological evaluation — in vitro cell-death assays across our cell-line panels; in vivo evaluation in cell-line-derived xenograft (CDX) models
• Compound evaluation — validation of research-use compounds and, where applicable, collaborative evaluation with industry partners

---

End-to-End Discovery Pipeline

The four foundations above are connected as a single, self-contained pipeline spanning five phases:

This integrated structure enables rapid iteration from hypothesis generation to experimental validation.

---

Disease-Focused Programs

We apply the four foundations above to cancers in which a defined genetic vulnerability is present but molecularly targeted treatment options remain limited. Our programs are organized in two categories.

Note: Mutation frequencies cited here reflect literature values; full citations are available on the Publications page and Research Highlights.

---

Category A — Rare, Pediatric, and AYA Cancers

Project A-1: SMARCB1-Deficient Rare Cancers — Malignant Rhabdoid Tumor and Epithelioid Sarcoma

---

Project A-2: ARID1A-Deficient Ovarian Clear Cell Carcinoma

---

Category B — Refractory Solid Tumors

Project B-1: Non-Small Cell Lung Cancer (NSCLC)

---

Project B-2: Diffuse-Type Gastric Cancer (Including Scirrhous Gastric Cancer)

---

Project B-3: Pancreatic Cancer and Esophageal Cancer

These programs are grouped as exploratory efforts centered on recurrent loss-of-function alterations and resistance-associated genetic contexts rather than a single histology.

---

Project B-4: PBRM1-Deficient Renal Cell Carcinoma

---

For Industry and Academic Partners

We actively pursue collaborations with pharmaceutical and academic partners, anchored in mechanism-driven target identification and a track record disclosed in peer-reviewed publications.

Our platform assets available for collaboration:

• A proprietary paralog-pair dual-knockout library — not represented in public dependency databases; enables recovery of paralog-buffered targets that single-gene screens cannot access.
• In-house rare- and refractory-cancer cell-line panels — including patient-derived cell lines for gastric and pancreatic cancers, and comprehensive panels spanning lung cancer, sarcoma, esophageal cancer, and renal cancer; covering cancer types and molecular subtypes underrepresented in international public databases.
• End-to-end mechanistic and pharmacological evaluation — from target identification, through mechanistic dissection, to pharmacological validation and indication-expansion analysis.

Published track record of industry collaboration:

The following is disclosed in the corresponding peer-reviewed publication:

• Ono Pharmaceutical Co., Ltd. × Division of Cancer Therapeutics — GCLC inhibitor program (GCLCi1 / GCLCi0) in SMARCB1-deficient rare cancers — Cancer Res 2026 (Takeuchi et al.)

Common modes of collaboration:

→ Contact us via the Contact and Access page, indicating that the inquiry concerns collaboration.

---

Related Pages

• Research Highlights — six thematic deep dives:
• Conventional Synthetic Lethality — 1-to-1 paralog dependency
• Paralog Co-Inhibition — 1-to-2 synthetic-lethality strategy
• Next-Generation Synthetic Lethality — higher-order synthetic lethality (2-to-n)
• Data-Driven Target Discovery — DepMap reanalysis and USP8 discovery
• Glutathione Metabolic Vulnerability — ARID1A and SMARCB1 metabolic targets
• Drug Repositioning — gemcitabine repositioning on a defined mechanism
• Publications — complete publication record and selected publications by research theme
• Lab Members — laboratory team and research environment
• Contact and Access — collaboration, academic, and media inquiries

---

Last Updated: 2026-05-19