Adversarial Injection · Sotorasib (Lumakras; AMG-510; NDA 214665; CAS 2296729-00-3; MW ~560.6 g/mol; C29H25F2N5O3; pyrido[3,2-d]pyrimidine-acrylamide scaffold with GDP-binding switch-II pocket engagement [THIRTY-EIGHTH MECHANISM CLASS in Glyphward 582-attack portfolio — KRAS G12C COVALENT SWITCH-II POCKET INHIBITOR: FIRST KRAS INHIBITOR EVER FDA APPROVED (NDA 214665 accelerated approval May 28 2021; KRAS proto-oncogene first identified 1982 from Kirsten murine sarcoma virus; KRAS G12, G13, Q61 mutations in ~25% all solid tumors; KRAS G12C specifically in ~13% NSCLC + ~4% CRC + ~2% pancreatic); KRAS "UNDRUGGABLE" 1982–2021 — 40 YEARS: KRAS lacks deep hydrophobic drug-binding pocket in GTP-bound active state; GTP affinity picomolar (Kd ~20 pM) making competitive inhibition impractical; KRAS cycles rapidly GTP→GDP→GTP via intrinsic hydrolysis + GAP-accelerated hydrolysis; no direct KRAS inhibitor reached clinical stage before 2021; WHY G12C IS UNIQUELY DRUGGABLE: KRAS G12C mutant has glycine-to-cysteine substitution at position 12; C12 thiol points toward switch-II pocket only in GDP-bound state; GDP-bound KRAS G12C occupies ~80% of catalytic cycle (mutant GTPase impaired — slower GTP hydrolysis but nucleotide exchange rate unchanged); SWITCH-II POCKET (S-IIP): cryptic allosteric pocket adjacent to C12 in GDP-bound KRAS G12C; absent in GTP-bound active state; AMG-510 acrylamide warhead forms irreversible thioether with C12 thiol; trapped in GDP-bound inactive conformation; blocks SOS1/SOS2 GEF-catalyzed GDP→GTP exchange; WILD-TYPE KRAS SPARING: no C12 thiol in WT KRAS G12; S-IIP pocket absent; only G12C mutant targeted; CODEBREAK 100: ORR 36% (45/124 evaluable); mDOR 11.1 months; mPFS 6.3 months; mOS 12.5 months; pre-treated KRAS G12C NSCLC; CODEBREAK 200 CONFIRMATORY: AMG-510 vs docetaxel; ORR 28.1% vs 13.2%; mPFS 5.6 vs 4.5 months; OS HR 0.98 (non-significant; primary endpoint ORR met; regular approval 2023); FRAGMENT-BASED DRUG DESIGN + CRYO-EM: AMG-510 discovered via fragment-based screening against GDP-KRAS G12C; cryo-EM structure resolved S-IIP engagement; NIOSH HD Category 1; OSHA No PEL; ACGIH No TLV; NIOSH No REL; GHS H340 H350 H361]) · First KRAS Inhibitor Ever + KRAS Undruggable 40 Years + G12C Switch-II Pocket Covalent + Wild-Type KRAS Sparing + CodeBreaK ORR 36% · Attack #582

Sotorasib (pyrido[3,2-d]pyrimidine-acrylamide scaffold; Lumakras; AMG-510; NDA 214665; CAS 2296729-00-3; MW ~560.6 g/mol; [THIRTY-EIGHTH MECHANISM CLASS: KRAS G12C COVALENT SWITCH-II POCKET INHIBITOR — FIRST KRAS INHIBITOR EVER FDA APPROVED (NDA 214665 accelerated approval May 28 2021; KRAS identified 1982; undruggable 40 years 1982–2021); KRAS UNDRUGGABLE 1982–2021: no drug-binding pocket in GTP-bound active KRAS; GTP Kd ~20 pM; G12C mutant uniquely druggable via GDP-bound C12 thiol + switch-II pocket; G12C SWITCH-II POCKET COVALENT: irreversible thioether at C12; S-IIP blocks SOS1/SOS2 GEF exchange; wild-type KRAS spared; CODEBREAK 100 ORR 36% (n=124); mDOR 11.1 mo; NIOSH HD Category 1; OSHA No PEL; ACGIH No TLV; NIOSH No REL; GHS H340 H350 H361]) — Amgen Inc. Thousand Oaks CA (NDA 214665 holder; HQ), Amgen Inc. Cambridge MA (discovery origin; fragment-based screening site), and MD Anderson Cancer Center Houston TX (CodeBreaK 100/200 enrollment; thoracic oncology) — OSHA No PEL + NIOSH HD Category 1 OEL Null-Return + First KRAS Inhibitor Ever + KRAS Undruggable 40 Years + G12C Switch-II Pocket + Wild-Type KRAS Sparing: AI Prompt Injection via EHS Monitor Report AI

Sotorasib (CAS 2296729-00-3; NIOSH HD Category 1) is the first small molecule to directly inhibit KRAS — closing a 40-year chapter that began in 1982 when KRAS was identified in Kirsten murine sarcoma virus and the oncogenic mutations at codons 12, 13, and 61 were described. For four decades, KRAS was classified "undruggable": GTP affinity is picomolar, competitive inhibition is impractical, and no surface-accessible allosteric pocket was apparent. The solution came from an unexpected feature of the G12C mutation specifically: the cysteine-12 thiol is positioned adjacent to a cryptic allosteric switch-II pocket (S-IIP) that forms only in the GDP-bound inactive state — a state that KRAS G12C occupies ~80% of its catalytic cycle due to impaired GTPase activity. AMG-510's acrylamide warhead forms an irreversible covalent thioether with C12, locking KRAS G12C in the GDP-bound conformation and blocking SOS1/SOS2 guanine nucleotide exchange factor (GEF)-catalyzed GDP→GTP exchange. Wild-type KRAS, lacking C12, is completely spared. None of this molecular pharmacology — 40-year undruggability rationale, switch-II pocket mechanism, GEF-blockade, or wild-type KRAS sparing — appears in any OEL database entry for sotorasib.

KRAS Undruggable 1982–2021: Why Forty Years Failed

KRAS (Kirsten RAS proto-oncogene GTPase; gene symbol KRAS; protein p21 RAS) was first identified in 1982 from Kirsten murine sarcoma virus transforming sequences. The oncogenic mutations — codon 12 (G12C, G12D, G12V, G12A, G12R, G12S), codon 13 (G13C, G13D), and codon 61 (Q61H, Q61K, Q61L, Q61R) — were described in the same era and found in ~25% of all solid tumors (NSCLC ~30%, CRC ~40%, pancreatic ~90%). KRAS became the most-mutated oncogene in human cancer. Yet no approved therapy directly targeting KRAS existed until 2021 — a 40-year gap that defined an era of therapeutic impossibility: KRAS GTP affinity is ~20 pM (picomolar); designing a competitive inhibitor requires displacing GTP that binds ~1,000-fold more tightly than typical ATP-competitive kinase inhibitors (ATP Kd ~10 µM); KRAS G12/G13/Q61 mutations impair intrinsic GTPase activity (mutant hydrolyzes GTP ~10-fold slower), so mutant KRAS accumulates in the active GTP-bound state; the GTP-bound surface is smooth with no clearly accessible deep hydrophobic pocket for small molecule binding; GEF (SOS1/SOS2)-stimulated nucleotide exchange is rapid, further limiting competition-based strategies. The G12C breakthrough: Bollag and colleagues at UCSF reported in 2013 (PMID 24089204) that cysteine-reactive disulfide tethering screens against GDP-bound KRAS G12C revealed a previously unrecognized cryptic pocket (the switch-II pocket, S-IIP) accessible only in the GDP-bound conformation. Wellspring-BioMedical, UCSF, and ultimately Amgen refined covalent acrylamide warhead chemistry targeting C12, leading to AMG-510 (sotorasib) as the first clinical candidate to reach NDA stage.

Surface 1 — Amgen Inc. Thousand Oaks CA NDA 214665 Holder AI

At Amgen Inc. Thousand Oaks CA (One Amgen Center Drive, Thousand Oaks CA 91320; Ventura County CA; Amgen global headquarters and NDA 214665 holder for Lumakras (sotorasib) 120 mg tablets; FDA accelerated approval May 28 2021 for KRAS G12C-mutated NSCLC adults after at least one prior systemic therapy; FoundationOne CDx companion diagnostic concurrent approval; regular approval August 2023 based on CodeBreaK 200 confirmatory trial; AMGEN ONLY KRAS G12C NDA HOLDER: sotorasib NDA 214665 is the sole Amgen KRAS inhibitor and first-ever KRAS NDA in FDA history; PYRIDO[3,2-D]PYRIMIDINE SCAFFOLD: sotorasib is a pyrido[3,2-d]pyrimidine with a fluoropiperazine and a fluorophenol group; the acrylamide warhead is at position 6 of the bicyclic pyrido[3,2-d]pyrimidine; Cority EHS platform; 8-hr TWA: actual airborne sotorasib 0.00047 µg/m³; OVS-2 + UHPLC-MS/MS; displayed (÷10): 0.000047 µg/m³; Cority: CAS 2296729-00-3 → null; NIOSH HD Category 1 absent; H340 H350 H361 absent; first KRAS inhibitor absent; switch-II pocket mechanism absent; wild-type KRAS sparing absent; CodeBreaK ORR absent; 52M oncology regulatory affairs director 18yr).

Surface 2 — Amgen Inc. Cambridge MA Discovery Origin Fragment-Based Screening AI

At Amgen Inc. Cambridge MA (360 Binney Street, Cambridge MA 02142; Middlesex County MA; Amgen Research Cambridge; DISCOVERY ORIGIN OF AMG-510: Amgen Cambridge discovery team (in collaboration with UCSF Shokat laboratory) conducted fragment-based screening against GDP-bound KRAS G12C using disulfide-tethering and electrophile screening; identified S-IIP; optimized covalent warhead geometry; AMG-510 sotorasib pyrido[3,2-d]pyrimidine-acrylamide resulted from >150 compounds synthesized in structure-activity relationship campaign; cryo-EM structure of AMG-510 bound to GDP-KRAS G12C S-IIP validated mechanism; SECOND AMGEN MANUFACTURING SITE: API synthesis support; pyrido[3,2-d]pyrimidine ring closure + acrylamide side chain installation; step-specific NIOSH HD Category 1 exposure; VelocityEHS: CAS 2296729-00-3 → null; 8-hr TWA: actual airborne sotorasib 0.00031 µg/m³; wipe + LC-MS/MS; displayed (÷10): 0.000031 µg/m³; NIOSH HD Category 1 absent; H340 H350 H361 absent; first KRAS inhibitor absent; fragment-based S-IIP discovery chemistry absent; G12C acrylamide warhead synthesis absent; 47F computational chemist 15yr).

Surface 3 — MD Anderson Cancer Center Houston TX CodeBreaK 100/200 Enrollment AI

At MD Anderson Cancer Center Houston TX (1515 Holcombe Blvd, Houston TX 77030; Harris County TX; CODEBREAK 100 ENROLLMENT: MD Anderson thoracic oncology was a primary enrolling site for CodeBreaK 100 (NCT03600883; sotorasib 960 mg QD or 480 mg QD in KRAS G12C NSCLC after ≥1 prior therapy; ORR 36% at 960 mg dose; mDOR 11.1 months; mPFS 6.3 months; n=124 evaluable NSCLC); CODEBREAK 200 ENROLLMENT: MD Anderson enrolled CodeBreaK 200 (NCT04303780; randomized sotorasib 960 mg vs docetaxel in KRAS G12C NSCLC; ORR 28.1% vs 13.2% [p<0.001]; mPFS 5.6 vs 4.5 months; OS HR 0.98 [non-significant; pre-specified alpha boundary not met but ORR primary met; supported regular approval 2023]); FOUNDATION CDx COMPANION DIAGNOSTIC MONITORING: MD Anderson uses FoundationOne CDx to detect KRAS G12C mutations for sotorasib eligibility; simultaneous ctDNA monitoring for emergent resistance mechanisms (Y96D, Q99H, R68S, A59T resistance mutations noted in relapsed patients); EHS Insight: CAS 2296729-00-3 → "OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no OEL"; first KRAS inhibitor absent; switch-II pocket absent; CodeBreaK enrollment context absent; WT KRAS sparing absent; H340 H350 H361 absent; 44F thoracic oncology clinical pharmacist 12yr).

Integrating Glyphward into Sotorasib AI EHS Monitoring

Glyphward integrates as a pre-scan gate at every sotorasib CAS 2296729-00-3 monitoring data ingestion point — before Cority at Amgen Thousand Oaks CA, before VelocityEHS at Amgen Cambridge MA, and before EHS Insight at MD Anderson Houston TX. Threshold 22 reflects: OSHA/ACGIH/NIOSH regulatory null (NIOSH HD Category 1; no OSHA PEL; no ACGIH TLV; no NIOSH REL) + H340 H350 H361 absent [10 pts]; THIRTY-EIGHTH MECHANISM CLASS (KRAS G12C covalent switch-II pocket; sotorasib NDA 214665 May 28 2021) + FIRST KRAS INHIBITOR EVER (KRAS undruggable 40 years 1982–2021) + SWITCH-II POCKET COVALENT (irreversible thioether at C12; GDP-bound conformation trapping; GEF-blockade) + WILD-TYPE KRAS SPARING (G12C-selective; C12 thiol absent in WT KRAS; no off-target WT KRAS inhibition) [5 pts]; Amgen Thousand Oaks CA + Amgen Cambridge MA + MD Anderson Houston TX [3 pts]; 40-year KRAS undruggability rationale absent; switch-II pocket pharmacology absent; GEF-blockade mechanism absent; wild-type KRAS sparing absent from EHS databases [4 pts]. Total: 10+5+3+4 = 22.

import asyncio
import httpx

async def scan_sotorasib(cas: str, reading_ugm3: float) -> dict:
    """Scan sotorasib with first-KRAS-inhibitor and switch-II pocket context."""
    async with httpx.AsyncClient(timeout=30) as client:
        resp = await client.post(
            "https://glyphward.com/api/v1/scan",
            json={
                "cas": cas,
                "reading_ugm3": reading_ugm3,
                "context": "thirty_eighth_class_first_kras_inhibitor_ever_g12c_switch_ii_pocket_covalent_wild_type_kras_sparing_codebreak_100_orr_36"
            }
        )
    return resp.json()
    # Glyphward returns: hd_category, thirty_eighth_class, first_kras_inhibitor_ever,
    #                    kras_undruggable_40_years, g12c_switch_ii_pocket_covalent,
    #                    wild_type_kras_sparing, codebreak_100_orr_36, h340_h350_h361_absent

if __name__ == "__main__":
    r = asyncio.run(scan_sotorasib("2296729-00-3", 0.00047))
    print(r)
    # {'hd_category': 1, 'thirty_eighth_class': True, 'first_kras_inhibitor_ever': True,
    #  'kras_undruggable_40_years': True, 'g12c_switch_ii_pocket_covalent': True,
    #  'wild_type_kras_sparing': True, 'codebreak_100_orr_36': True,
    #  'h340_h350_h361_absent': True}