Esophageal Cancer: GWAS to Drug Target Druggability Analysis

COMPREHENSIVE GWAS-TO-DRUG-TARGET DRUGGABILITY ANALYSIS: ESOPHAGEAL CANCER

Section 1: Disease Identifiers

Notes: MONDO:0007576 is the primary entry for esophageal cancer susceptibility; MONDO:0019086 covers the broader carcinoma grouping. Both map to overlapping but distinct GWAS study sets. The disease has extensive clinical trial coverage (1,335 trials linked from MONDO:0007576).

Section 2: Gwas Landscape

Summary:

• Total GWAS associations: 101 (52 from MONDO:0007576 + 49 from EFO:0002916)

• Unique studies: ~30 (GCST000466, GCST000777, GCST001089, GCST001672, GCST001674, GCST002010, GCST90011807, GCST90013699, GCST90018621, GCST90018841, GCST90308755, GCST90308764, GCST90651054, GCST90651060, GCST90651069, GCST90651166, and others)

• Unique mapped genes: ~55

TOP 50 GWAS ASSOCIATIONS (ranked by p-value):

Section 3: Variant Details

Functional Classification of Top GWAS Loci:

The GWAS Catalog associations for esophageal cancer predominantly implicate non-coding regions, with key exceptions at alcohol metabolism loci. ADH1B rs1229984 (His48Arg) is a well-characterized missense variant (Tier 1) — the strongest signal in the entire dataset. ALDH2 rs671 (Glu504Lys) is also a missense variant — the second-strongest signal. Both are coding variants in enzyme active sites.

MAF Distribution: The ADH1B His48Arg and ALDH2 Glu504Lys variants show dramatic population-specific frequency differences — rare in Europeans (<5%) but common in East Asians (30-40%), consistent with the high incidence of esophageal squamous cell carcinoma in East Asian populations.

Summary:

• Coding variants: 3 (6%)
• Splice/UTR: 4 (8%)
• Regulatory: 28 (56%)
• Intronic: 15 (30%)

Section 4: Mendelian Disease Overlap

GenCC curated gene-disease relationships for esophageal cancer (MONDO:0007576):

ClinVar pathogenic variants in esophageal cancer genes (MONDO:0007576):

HPO-linked genes (HP:0011459 - Esophageal carcinoma):

Genes with BOTH GWAS + Mendelian evidence = HIGHEST CONFIDENCE:

• CHEK2 — GWAS p=2e-22 in ESCC + Mendelian cancer susceptibility gene
• PLCE1 — GWAS p=4e-20 + linked to gastric/esophageal cancer susceptibility
• CASP8 — GWAS p=1e-8 + cancer susceptibility (breast, melanoma, basal cell)

Section 5: Gwas Genes To Proteins

Summary: 55 unique gene loci → ~42 protein-coding genes with UniProt entries

TOP 50 GWAS GENES MAPPED TO PROTEINS:

Section 6: Protein Family Classification

Summary:

Section 7: Expression Context

Disease-relevant tissues: Esophageal epithelium (stratified squamous), gastric cardia (for adenocarcinoma subtype), liver (alcohol metabolism).

Based on biobtree Bgee expression data and known biology:

Key findings:

• PSCA, ECRG4, RHBDF2 show esophageal tissue-enriched expression — therapeutic targeting would have fewer off-target effects
• KCNQ1 has strong cardiac expression — targeting requires careful cardiac safety monitoring
• ADH1B/ALDH2 are liver/GI enriched, consistent with alcohol metabolism pathway

Section 8: Protein Interactions

ERBB2 (HER2) interaction network (STRING, score >900):

ERBB2 is a major hub with 7,626 interactions. Top interactors include:

• EGFR (P00533, score 983) — DRUGGED (cetuximab, erlotinib, gefitinib)
• ERBB3 (P21860, score 984) — DRUGGED (pertuzumab targets ERBB2/3)
• ERBB4 (Q15303, score 983) — DRUGGED (lapatinib)
• SRC (P12931, score 980) — DRUGGED (dasatinib, bosutinib)
• HSP90AA1 (P07900, score 985) — DRUGGED (ganetespib)
• PIK3CA (P42336, score 927) — DRUGGED (alpelisib)
• TP53 (P04637, score 924) — PPI inhibitors
• BRCA1 (P38398, score 919) — PARP synthetic lethality
• BRCA2 (P51587, score 874) — PARP synthetic lethality
• ESR1 (P03372, score 966) — DRUGGED (tamoxifen)

Undrugged GWAS genes interacting with drugged genes:

Section 9: Structural Data

PDB Structure Availability for Key GWAS Proteins:

Undrugged targets with structure:

Summary: 22 proteins with PDB structures / 10 AlphaFold only / 0 with no structural data

Section 10: Drug Target Analysis

ChEMBL Drug Target Summary:

Proteins with approved drugs (Phase 4):

Approved drugs for OTHER cancers with GWAS gene targets:

OPPORTUNITY GAP:

• 20 GWAS genes (48%) have NO drug development at all — including PLCE1 (p=4e-20), RUNX1 (p=8e-22), HECTD4 (p=2e-31), HEATR3 (p=2e-20), CUX2, TOX3, RHBDF2, DLEC1, WWOX, RNF6

Section 11: Bioactivity & Enzyme Data

TOP GWAS proteins by ChEMBL bioactivity data:

Enzyme GWAS genes (druggability assessment):

Undrugged genes with bioactivity starting points:

• PSCA (O43653, CHEMBL3712961) — Prostate stem cell antigen, GPI-anchored surface protein → antibody target (ADC approach)
• ACAD10 (Q6JQN1, CHEMBL4105816) — Acyl-CoA dehydrogenase → limited compound data
• MSR1 (P21757, CHEMBL5811) — Scavenger receptor → some small-molecule modulators
• SLC39A6 (Q13433, CHEMBL4523581) — Zinc transporter → limited chemical matter

Section 12: Pharmacogenomics

All key GWAS genes have PharmGKB entries (all marked as VIP genes):

Section 13: Clinical Trials

Clinical trial landscape for Esophageal Cancer (MONDO:0007576):

• Total trials linked: 1,335+
• From the first 100 retrieved:

TOP 30 Drugs in Esophageal Cancer Trials:

GWAS gene alignment:

• 7 of 30 top trial drugs (23%) directly target GWAS/Mendelian genes (ERBB2, TP53, TGFBR2, FGFR2, BRCA2)
• Majority of trials (77%) use cytotoxics or immune checkpoint inhibitors — not genetically informed
• Significant disconnect between genetic evidence and clinical development

Section 14: Pathway Analysis

Key Reactome Pathways Enriched Among GWAS Genes:

Pathway-level druggability insight: Even where GWAS genes themselves are undrugged, their pathway partners often are. For example:

• RUNX1 (undrugged TF) → HDAC pathway → entinostat, tucidinostat
• SMAD7 (undrugged) → TGF-beta pathway → galunisertib, vactosertib
• HECTD4 (undrugged E3 ligase) → ubiquitin-proteasome → bortezomib, carfilzomib
• XBP1 (undrugged TF) → UPR/IRE1 pathway → IRE1 inhibitors
• STAT1 (difficult TF) → JAK-STAT → ruxolitinib, tofacitinib

Section 15: Drug Repurposing Opportunities

Prioritized Repurposing Candidates:

Scoring criteria: Genetic evidence (0-3) + Mendelian overlap (0-2) + Druggable family (0-2) + Tissue expression (0-1.5) + Safety profile (0-1.5) = max 10

Section 16: Druggability Pyramid

Section 17: Undrugged Target Profiles

TOP 30 High-Value Undrugged Targets (ranked by druggability potential):

1. PLCE1 — Phospholipase C Epsilon 1

• GWAS p-value: 4e-20 | Variant type: Regulatory
• Protein function: Phospholipase C enzyme; hydrolyzes PIP2 to DAG and IP3; also has RasGEF activity
• Family: Enzyme (phospholipase) — DRUGGABLE
• Structure: 3 PDB (RA domains only), AlphaFold pLDDT=61
• Expression: Esophagus, stomach, heart, blood vessels
• Interactions: RAS family, PKC pathway
• Why undrugged: Large multi-domain protein (2,303 aa); no selective inhibitors developed
• Druggability potential: MEDIUM-HIGH — enzymatic activity is druggable in principle; needs structure-based design

2. HECTD4 — HECT Domain E3 Ubiquitin Ligase 4

• GWAS p-value: 2e-31 | Variant type: Intronic
• Protein function: E3 ubiquitin ligase; protein ubiquitination
• Family: E3 ligase — EMERGING (PROTAC/molecular glue potential)
• Structure: No PDB, AlphaFold available
• Expression: Ubiquitous
• Interactions: Ubiquitin-proteasome pathway
• Why undrugged: Novel target; HECT domain potentially druggable
• Druggability potential: MEDIUM — E3 ligase hijacking (PROTAC) is emerging modality

3. RUNX1 — Runt-Related Transcription Factor 1

• GWAS p-value: 8e-22 | Variant type: Intronic
• Protein function: Transcription factor; master regulator of hematopoiesis
• Family: Transcription factor — DIFFICULT
• Structure: PDB structures available for Runt domain
• Expression: Hematopoietic cells, esophagus
• Interactions: HDAC1/2, CBFβ
• Why undrugged: TF — no binding pocket for small molecules
• Druggability potential: LOW (but HDAC pathway druggable)

4. RHBDF2 — Inactive Rhomboid Protein 2 (iRhom2)

• GWAS p-value: Not in GWAS (Mendelian — tylosis with esophageal cancer)
• Protein function: Regulates ADAM17 sheddase; controls TNF-alpha shedding
• Family: Inactive protease / regulatory protein
• Structure: 5 PDB (ADAM17 complex, cryo-EM, 2.3 Å)
• Expression: Skin, esophagus — HIGH RELEVANCE
• Interactions: ADAM17 (drugged partner — MEDI3622 antibody)
• Why undrugged: Recently characterized; structure only available since 2023
• Druggability potential: HIGH — ADAM17 complex structures enable drug design; directly causative of familial esophageal cancer

5. WWOX — WW Domain Oxidoreductase

• GWAS p-value: ClinVar evidence
• Protein function: Tumor suppressor; short-chain dehydrogenase/reductase
• Family: SDR enzyme — DRUGGABLE
• Structure: 1 PDB (WW2 domain), AlphaFold pLDDT=85
• Expression: Brain, ovary, prostate, esophagus
• Interactions: p53, ERBB4, AP2gamma
• Why undrugged: Tumor suppressor — therapeutic would need to restore/enhance function
• Druggability potential: MEDIUM — SDR domain potentially druggable; activator approach needed

6. PSCA — Prostate Stem Cell Antigen

• GWAS p-value: 3e-13 | Variant type: Regulatory
• Protein function: GPI-anchored cell surface protein
• Family: Ly-6 superfamily — SURFACE ANTIGEN
• Expression: Prostate, esophagus, stomach, bladder — HIGH RELEVANCE
• Why undrugged: Surface antigen — ideal for antibody/ADC/CAR-T approach
• Druggability potential: HIGH — cell surface target; ADC and bispecific antibody approaches feasible

7. ECRG4 — Esophageal Cancer Related Gene 4

• GWAS p-value: 1e-8 | Variant type: Regulatory
• Protein function: Secreted peptide (augurin); anti-inflammatory, anti-proliferative
• Family: Secreted peptide
• Expression: Esophagus-enriched — HIGHEST RELEVANCE
• Why undrugged: Tumor suppressor (peptide) — recombinant protein approach
• Druggability potential: MEDIUM — replacement therapy or pathway activation

8-10. DCC, DLEC1, RNF6

• DCC: Netrin receptor; dependence receptor — anti-netrin antibodies in clinical development
• DLEC1: Deleted in lung/esophageal cancer — tumor suppressor, re-expression strategy
• RNF6: E3 ubiquitin ligase — emerging druggable class

11-20. HEATR3, CUX2, TOX3, HNF1B, XBP1, MAF, SMAD7, LZTS1, CTHRC1, ASCC1

• Mostly transcription factors or scaffold proteins — difficult targets
• XBP1 is notable: UPR pathway member → IRE1 inhibitors available preclinically
• SMAD7: TGF-beta pathway → druggable pathway partners (galunisertib, vactosertib)

Section 18: Summary

GWAS LANDSCAPE

• Total associations: 101 across ~30 GWAS studies
• Unique genes: 42 protein-coding genes
• Coding vs non-coding variants: 6% coding / 94% non-coding
• Strongest signals: ADH1B (p=2e-46), HECTD4 (p=2e-31), ALDH2 (p=1e-25), FGFR2 (p=1e-25), CHEK2 (p=2e-22)

GENETIC EVIDENCE

• Tier 1 (coding) genes: 3 (ADH1B, ALDH2, CHEK2)
• Mendelian overlap genes: 14 (RHBDF2, ERBB2, TGFBR2, TP53, BRCA2, DCC, STAT1, WWOX, DLEC1, RNF6, MSR1, LZTS1, CTHRC1, ASCC1)
• Both GWAS + Mendelian: 3 (CHEK2, CASP8, PLCE1)

DRUGGABILITY

• Overall druggable rate: 52% of GWAS genes have druggable protein families
• Approved drugs (for any disease): 12 genes (29%)
• Approved FOR esophageal cancer: 2 genes (5%) — ERBB2, TP53
• In clinical trials: 4 genes (10%)
• Opportunity gap (no drug development): 18 genes (43%)

PYRAMID SUMMARY

CLINICAL TRIAL ALIGNMENT

• 23% of top trial drugs target GWAS genes — moderate alignment
• Most trials focus on immune checkpoint inhibitors (PD-1/PD-L1) and cytotoxics — NOT genetically informed
• FGFR2 inhibitors are a major missed opportunity — approved for other cancers, strong GWAS signal (p=1e-25) in esophageal cancer

TOP 10 REPURPOSING CANDIDATES

TOP 10 UNDRUGGED OPPORTUNITIES

TOP 10 INDIRECT OPPORTUNITIES

KEY INSIGHTS

1. Alcohol metabolism dominates the genetic architecture: ADH1B and ALDH2 are the strongest GWAS signals (p=2e-46 and 1e-25), both coding variants in enzyme active sites. This is unique to esophageal cancer — particularly squamous cell carcinoma in East Asian populations where these variants are common.

2. FGFR2 is the top repurposing opportunity: GWAS evidence (p=1e-25), multiple approved inhibitors for other cancers (erdafitinib, pemigatinib, futibatinib), excellent structural data, and esophageal tissue expression — yet no FGFR2-targeted trials specifically for esophageal cancer are in the top listings.

3. RHBDF2 is the most compelling undrugged target: Causative gene for tylosis with esophageal cancer (Mendelian), directly implicated via ADAM17 sheddase complex, and recent cryo-EM structures (2023) enable structure-based drug design. This represents the strongest genetic-to-functional link in the disease.

4. Immunotherapy-genetics disconnect: The clinical trial landscape is dominated by immune checkpoint inhibitors (PD-1/PD-L1), yet the genetic architecture points to alcohol metabolism (ADH1B/ALDH2), receptor tyrosine kinase signaling (FGFR2/ERBB2), and DNA damage response (CHEK2/BRCA2) pathways. Only PTPN2, which emerged as an immuno-oncology target (enhancing anti-tumor immunity when inhibited), bridges these worlds.

5. PTPN2 as an emerging opportunity: GWAS evidence in ESCC (p=2e-11), plus strong evidence in autoimmune diseases, makes PTPN2 a compelling immuno-oncology target. ABBV-CLS-484 (PTPN2 degrader) is in Phase 1 trials for solid tumors — esophageal cancer should be considered for basket trial inclusion.

6. Compared to other cancers: Esophageal cancer’s genetic architecture is unusual in being dominated by metabolic enzyme variants (ADH1B, ALDH2) rather than classical oncogenes/tumor suppressors. The 8q24 locus (PCAT1/CASC8/POU5F1B, p=9e-29) is shared with colorectal, prostate, and breast cancers. The FGFR2 locus is shared with breast cancer (p=2e-76).

7. E3 ligase opportunity: HECTD4 (p=2e-31) and RNF6 (GenCC) are E3 ubiquitin ligases — an emerging druggable class via molecular glue degraders (analogous to lenalidomide/CRBN). The PROTAC modality could unlock these targets.

Analysis performed using biobtree MCP tools querying 70+ biological databases including GWAS Catalog, MONDO, EFO, MeSH, HPO, OMIM, Orphanet, GenCC, ClinVar, UniProt, ChEMBL, InterPro, PDB, AlphaFold, STRING, Reactome, PharmGKB, Bgee, and Guide to Pharmacology. 2026-04-11.