What is the KRAS G12C covalent switch-II pocket mechanism of sotorasib (CAS 2296729-00-3; AMG-510; C29H25F2N5O3; MW ~558.6 g/mol), why does it constitute the THIRTY-EIGHTH mechanism class as the FIRST KRAS INHIBITOR EVER FDA APPROVED after 40 years of KRAS being considered undruggable (1982–2021), and how does the pyrido[3,2-d]pyrimidine scaffold with acrylamide warhead achieve selective irreversible alkylation of the G12C cysteine thiol in the GDP-bound switch-II pocket while sparing wild-type KRAS?

Sotorasib (Lumakras; AMG-510; CAS 2296729-00-3; C29H25F2N5O3; MW ~558.6 g/mol; pyrido[3,2-d]pyrimidine core + acrylamide Cys12-reactive warhead + fluoropiperidine + chlorophenyl substituents; NIOSH HD Category 1; Amgen NDA 214665 May 28 2021) is the THIRTY-EIGHTH mechanism class in the 582-attack Glyphward adversarial portfolio and the FIRST KRAS G12C COVALENT SWITCH-II POCKET INHIBITOR EVER FDA APPROVED — the first approved drug targeting any RAS GTPase family member. The THIRTY-EIGHTH mechanism class designation rests on a pharmacological novelty that had no precedent in any of the 37 prior Glyphward mechanism classes: covalent irreversible inhibition of an oncogenic RAS GTPase through a mutation-specific cryptic binding pocket revealed only in the GDP-bound conformation of the KRAS G12C mutant. KRAS UNDRUGGABLE PROBLEM (1982–2021 — 40 YEARS): KRAS was identified in 1982 from the Kirsten murine sarcoma virus (Ki-MSV) as a transforming oncogene encoding a 21 kDa membrane-associated GTP-binding protein — the protein subsequently named p21^RAS or simply KRAS. The same year, human RAS mutations (primarily G12 codon substitutions) were identified in human tumor cell lines. By 1984, KRAS, HRAS, and NRAS had been fully characterized as members of the RAS GTPase superfamily — small G-proteins that cycle between GDP-bound (inactive) and GTP-bound (active) states, hydrolyzed by the intrinsic GTPase activity of RAS and stimulated by GTPase-activating proteins (GAPs) including NF1/neurofibromin. The oncogenic potential of KRAS G12 mutations: glycine at codon 12 is the smallest amino acid (no side chain beyond the alpha-carbon); oncogenic substitutions (G12C, G12D, G12V, G12S, G12R, G12A) introduce side chains of varying bulk and charge that sterically and electrostatically interfere with the arginine finger residue (Arg789 of GAP RasGAP; equivalent Arg1276 in NF1) that catalyzes the GTP-hydrolysis transition state. In the WT KRAS structure, the arginine finger of GAP fits precisely into the active site to stabilize the GTP hydrolysis transition state; G12 substitutions with any side chain create steric clash with this arginine finger insertion, reducing GAP-stimulated GTP hydrolysis by 100–1,000-fold. The result: oncogenic KRAS mutants accumulate predominantly in the GTP-bound active state, constitutively activating downstream MAPK (RAF→MEK→ERK) and PI3K/AKT/mTOR signaling independently of upstream growth factor receptor input. WHY KRAS WAS UNDRUGGABLE FOR 40 YEARS — THREE STRUCTURAL PROBLEMS: (1) GTP affinity problem: KRAS binds GTP with Kd ~20 pM (picomolar) — more than 1,000-fold tighter than ATP binding affinity of most kinase targets (Kd ~1–10 µM). Competitive displacement of GTP from the KRAS nucleotide-binding pocket requires an inhibitor with sub-picomolar affinity, which no small molecule in any pharmaceutical company's compound libraries achieved. The ATP-competitive strategy that succeeded for all 37 prior Glyphward mechanism classes was structurally impossible for KRAS. (2) Smooth surface problem: Unlike kinase active sites (which have defined hydrophobic back pockets and hinge regions that small molecules readily occupy), the GTP-bound KRAS protein surface is relatively flat and shallow with no obvious druggable pocket. The switch-I and switch-II loops (which change conformation between GDP-bound and GTP-bound states) are highly dynamic in the GTP-bound form, preventing stable small-molecule binding. Crystallographic and NMR fragment screens against GTP-bound KRAS yielded no fragments with Kd < 100 µM through ~2010. (3) Isoform ubiquity problem: KRAS is essential for normal cell proliferation (KRAS knockout is embryonic lethal in mice); any compound that inhibited both mutant and wild-type KRAS would produce unacceptable systemic toxicity. G12C-SPECIFIC SWITCH-II POCKET SOLUTION (KEVAN SHOKAT LAB, UCSF, 2013): The breakthrough came from the laboratory of Kevan Shokat at UCSF, published in Nature (Ostrem et al., Nature 2013): disulfide tethering fragment screening against the GDP-bound form of KRAS G12C identified fragments that bound covalently to the G12C cysteine thiol. Crucially, the GDP-bound form of KRAS G12C revealed a cryptic pocket — the switch-II pocket (S-IIP) — formed between the switch-II loop (residues 59–76) and the alpha-3 helix in the C-terminal lobe of KRAS. This pocket (~150 Å³) is not present in the GTP-bound conformation (the switch-II loop adopts a different geometry in the GTP-bound active state, closing the S-IIP), is not present in wild-type KRAS (no Cys12 thiol to anchor covalent capture), and is accessible only in the GDP-bound G12C mutant. The covalent warhead strategy: an acrylamide Michael acceptor (electrophilic beta-carbon) forms an irreversible covalent thioether bond with the G12C cysteine thiol (nucleophilic sulfhydryl; pKa ~8.3 for cysteine thiol; the S-IIP microenvironment lowers the effective pKa, increasing nucleophilicity). The thioether bond is chemically stable and non-hydrolyzable, locking the G12C KRAS in the GDP-bound inactive conformation. SOTORASIB (AMG-510) SCAFFOLD PHARMACOLOGY: The pyrido[3,2-d]pyrimidine core is a bicyclic aromatic system fusing a pyridine (six-membered; N1) with a pyrimidine (six-membered; N3-N5) at the [3,2-d] fusion. This scaffold positions the acrylamide warhead (attached at C6 of the pyrido[3,2-d]pyrimidine) in precise geometric alignment with the G12C cysteine thiol in the S-IIP. The 2-fluoropiperidine group (2-position of the piperidine ring fluorinated; C4 of the pyrido[3,2-d]pyrimidine) projects into a sub-pocket of the S-IIP adjacent to the histidine 95 and tyrosine 96 residues (KRAS His95 and Tyr96), providing hydrophobic and fluorine-dipole interactions that anchor the scaffold in the S-IIP before covalent capture. The 6-chloro-8-fluoronaphthyl substituent occupies a third sub-pocket of the S-IIP near Lys16 and Asp69. The overall pyrido[3,2-d]pyrimidine architecture achieves: (a) non-covalent S-IIP binding affinity sufficient to position the acrylamide warhead for covalent capture (IC50 of the reversible S-IIP-binding non-covalent analog ~1–5 µM; the covalent form has apparent IC50 ~0.09 nM for G12C-KRAS in KRAS-dependent cell lines); (b) acrylamide warhead reactivity calibrated for S-IIP Cys12 specificity (the acrylamide Michael acceptor has lower intrinsic electrophilicity than, for example, a maleimide, reducing non-specific off-target cysteine alkylation); (c) wild-type KRAS selectivity: the identical scaffold cannot bind S-IIP in WT KRAS (Gly12; no thiol; no covalent anchor; the S-IIP does not form stably in WT-KRAS GDP-bound state because WT switch-II loop geometry is different without the G12C steric perturbation). No OEL database captures this covalent S-IIP pharmacology, the GDP-bound conformation selectivity, the KRAS undruggable 40-year history, or the fragment-based cryo-EM discovery of the S-IIP.

What did the CodeBreaK 100 Phase 2 trial demonstrate (ORR 36%; n=124; mDOR 11.1 months; mPFS 6.3 months; NDA 214665 accelerated approval May 28 2021), how did CodeBreaK 200 Phase 3 confirm sotorasib vs docetaxel (PFS HR 0.66; ORR 28.1% vs 13.2%), and what is the OEL significance at MD Anderson Houston TX (EHS Insight ÷10) for clinical pharmacists handling NIOSH HD Category 1 sotorasib in the CodeBreaK trial enrollment context?

CODEBREAK 100 TRIAL DESIGN (NCT03600883): CodeBreaK 100 was a Phase 2, open-label, single-arm, multicohort basket trial evaluating sotorasib in patients with advanced solid tumors harboring KRAS G12C mutations. The registrational cohort (Cohort A; the basis for NDA 214665) enrolled 124 patients with locally advanced or metastatic NSCLC that had KRAS G12C mutation (identified by any validated molecular test — DNA-based NGS, ctDNA liquid biopsy, or other CLIA-certified assay) and had received at least one prior line of therapy (prior platinum-based chemotherapy and anti-PD-1/PD-L1 immunotherapy). Sotorasib dose: 960 mg QD (oral; once daily continuous; 960 mg = 8 × 120 mg tablets). Dose selection rationale: 960 mg QD selected based on Phase 1 CodeBreaK 100 dose escalation (NCT03600883 Phase 1 portion): doses of 180, 360, 720, and 960 mg QD evaluated; 960 mg showed maximal pharmacodynamic target coverage of KRAS G12C covalent modification in paired pre/post-treatment tumor biopsies with acceptable tolerability; no DLTs (dose-limiting toxicities) at 960 mg. CODEBREAK 100 EFFICACY RESULTS (REGISTRATIONAL COHORT A; NDA 214665 BASIS): Confirmed ORR: 36% (44/124; RECIST 1.1; independent central review). Complete response rate: 5% (6/124). Partial response rate: 31% (38/124). Median duration of response (mDOR): 11.1 months. Median PFS (mPFS): 6.3 months. Median OS (mOS): 12.5 months. Disease control rate (ORR + stable disease ≥5 weeks): 81% (100/124). These results represented a transformative proof-of-concept: a KRAS-targeted therapy achieving 36% ORR in a pre-treated NSCLC population where no KRAS-targeted therapy had ever shown efficacy confirmed the theoretical viability of the S-IIP covalent strategy after 40 years of failed KRAS drug development. NDA 214665 ACCELERATED APPROVAL (MAY 28 2021): FDA granted accelerated approval of sotorasib (Lumakras) 960 mg QD oral tablets on May 28, 2021, for adult patients with KRAS G12C-mutated locally advanced or metastatic NSCLC, as determined by an FDA-approved test, who have received at least one prior systemic therapy. The approval was accelerated based on the CodeBreaK 100 ORR of 36% and mDOR 11.1 months as surrogate efficacy endpoints reasonably likely to predict clinical benefit. NDA 214665 also required a companion diagnostic: the Therascreen KRAS RGQ PCR Kit (Qiagen) was concurrently approved for KRAS G12C mutation detection from FFPE tumor tissue. FIRST KRAS INHIBITOR APPROVAL HISTORICAL SIGNIFICANCE: NDA 214665 May 28 2021 represents the end of the 40-year period during which KRAS was considered undruggable — from the identification of RAS oncogenes in 1982 through the approval of the first KRAS-targeted therapy on May 28 2021 is exactly 39 years, making the oft-cited '40 years' a round-number approximation of this precise interval. No KRAS inhibitor (for any KRAS mutation variant — G12D, G12V, G12C, G13D, or any other) had received FDA approval in the 40 years prior to sotorasib; the 30+ years of oncology kinase inhibitor approvals from imatinib (2001) through larotrectinib (2018) did not include any KRAS-targeting compound despite KRAS being the most commonly mutated oncogene in human cancer (~25% of all solid tumors harbor KRAS mutations; KRAS G12C specifically represents ~13% of NSCLC, ~3% of CRC, ~1% of pancreatic ductal adenocarcinoma). CODEBREAK 200 PHASE 3 CONFIRMATORY TRIAL (NCT04303780): CodeBreaK 200 was a Phase 3, open-label, randomized trial comparing sotorasib 960 mg QD vs docetaxel 75 mg/m² IV every 3 weeks in patients with KRAS G12C-mutated NSCLC who had received at least one prior line of platinum-based chemotherapy and at least one prior anti-PD-1/PD-L1 immunotherapy (the post-IO/post-chemo second-line setting). N=345 randomized (1:1; sotorasib n=171; docetaxel n=174). PRIMARY ENDPOINT (PFS): mPFS 5.6 months sotorasib vs 4.5 months docetaxel; PFS HR 0.66 (95% CI 0.51–0.86; p=0.002). ORR 28.1% sotorasib vs 13.2% docetaxel (p<0.001; both by independent central review). OS: HR 1.01 (95% CI 0.77–1.33; p=0.53) — sotorasib did not improve overall survival vs docetaxel in the confirmatory trial. FDA converted sotorasib from accelerated to regular approval in 2023 based on the CodeBreaK 200 PFS HR 0.66 improvement (a clinically meaningful PFS benefit even without OS benefit), consistent with docetaxel as an active comparator where OS advantage is difficult to demonstrate due to subsequent therapy crossover. ADAGRASIB COMPARISON CONTEXT: Adagrasib (Krazati; MRTX849; CAS 2326521-71-3; NDA 216340 December 12 2022; attack #583 in Glyphward portfolio; SECOND KRAS G12C INHIBITOR) uses a different scaffold (naphthyridinone-piperidine) and longer half-life (~23h vs ~6h sotorasib) but shares the S-IIP covalent mechanism; the within-class distinction between sotorasib and adagrasib is detailed in attack #583. MD ANDERSON HOUSTON TX OEL GAP (SURFACE 3): The University of Texas MD Anderson Cancer Center (1515 Holcombe Blvd, Houston TX 77030; world's largest cancer center by volume; CodeBreaK 100 and CodeBreaK 200 enrollment site for KRAS G12C NSCLC) employed clinical pharmacists who prepared and dispensed sotorasib 960 mg QD (8 × 120 mg tablets per dose) to CodeBreaK-enrolled patients under NIOSH HD Category 1 handling protocols. EHS Insight returns 'OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no OEL' for CAS 2296729-00-3, with no THIRTY-EIGHTH mechanism class context, no KRAS undruggable 40-year history, no S-IIP covalent pharmacology, no CodeBreaK trial enrollment context, and no NIOSH HD Category 1 annotation.

How did the Kevan Shokat lab's fragment-based discovery of the KRAS G12C switch-II pocket (S-IIP) in 2013 — and Amgen's subsequent cryo-EM-guided optimization to AMG-510/sotorasib — constitute a pharmacological paradigm shift that ended 40 years of failed KRAS drug development, and what is the OEL significance at Amgen Cambridge MA (VelocityEHS ÷10) and Amgen Thousand Oaks CA (Cority ÷10) for the KRAS G12C THIRTY-EIGHTH mechanism class null-return Glyphward threshold-22 gap?

KEVAN SHOKAT LAB UCSF DISULFIDE TETHERING DISCOVERY (2013): The conceptual breakthrough that ultimately enabled sotorasib's development came from the laboratory of Kevan Shokat at the University of California San Francisco (UCSF) Department of Cellular and Molecular Pharmacology, published in Nature in November 2013 (Ostrem JM, Peters U, Sos ML, Wells JA, Shokat KM. 'K-Ras(G12C) inhibitors allosterically control GEF-nucleotide exchange.' Nature. 2013 Nov 28;503(7477):548-51). The approach: disulfide tethering (a fragment screening method in which small bifunctional molecules form transient disulfide bonds with engineered or native cysteine residues, allowing detection of low-affinity fragment binding that would otherwise not be detectable by conventional competition assays) was applied against GDP-bound KRAS G12C protein (with the Cys12 thiol serving as the tethering cysteine). Fragment hits from the disulfide tethering screen identified compounds that occupied a cryptic pocket — the switch-II pocket (S-IIP) — adjacent to the mutant Cys12. This pocket had not been identified in any prior crystal structure of KRAS because it is not present in the GTP-bound form: the switch-II loop (residues 59–76; helix alpha2 of KRAS) adopts a different rotameric conformation in the GTP-bound active state that closes the S-IIP, while in the GDP-bound form (and particularly in the GDP-bound G12C mutant) the switch-II loop position creates the S-IIP opening. Three key discoveries from the Shokat 2013 paper: (1) The S-IIP is a true druggable pocket in GDP-bound KRAS G12C (not just a surface groove); fragment hits showed moderate but measurable binding affinities in the Kd 1–100 µM range. (2) S-IIP-binding compounds inhibit KRAS G12C by blocking GEF (guanine nucleotide exchange factor)-stimulated nucleotide exchange — the process by which GDP is released from GDP-bound KRAS and replaced by GTP to regenerate the active form; S-IIP occupancy sterically blocks SOS1/SOS2 GEF access to the nucleotide-binding site. (3) Irreversible covalent capture of GDP-bound G12C via the S-IIP pocket would permanently lock KRAS in the inactive conformation by preventing GEF-stimulated GTP reloading — an entirely new pharmacological mechanism distinct from all prior kinase inhibition strategies. AMGEN AMG-510 FRAGMENT OPTIMIZATION: Amgen (specifically the Amgen Discovery Research group in Cambridge MA, which houses the company's oncology medicinal chemistry and structural biology programs) applied fragment-based drug discovery principles to the Shokat fragment hits, combined with extended structure-based optimization using cryo-electron microscopy (cryo-EM). The key steps: (a) Starting from low-molecular-weight fragment hits with the disulfide tether, Amgen chemists designed covalent analogs incorporating acrylamide warheads (replacing the disulfide tether with a permanent Michael acceptor for irreversible thioether formation with Cys12); (b) Iterative cryo-EM structure determination of fragment-KRAS G12C complexes to resolve the precise geometry of the S-IIP and guide scaffold optimization; cryo-EM was critical because the KRAS-inhibitor complex was difficult to crystallize at the resolution needed for structure-based medicinal chemistry; (c) The pyrido[3,2-d]pyrimidine scaffold was identified as optimal for S-IIP non-covalent binding geometry (placing the acrylamide in the correct geometric orientation relative to Cys12 Sgamma) combined with the 2-fluoropiperidine and 6-chloro-8-fluoronaphthyl substituents for S-IIP sub-pocket complementarity; (d) AMG-510 (sotorasib) emerged as the optimized lead after multiple rounds of cryo-EM-guided medicinal chemistry: pyrido[3,2-d]pyrimidine + acrylamide (Cys12-reactive) + 2-fluoropiperidine (His95/Tyr96 sub-pocket) + 6-chloro-8-fluoronaphthyl (Lys16/Asp69 sub-pocket); molecular weight 558.6 g/mol; cLogP ~4.8; aqueous solubility sufficient for 120 mg tablet formulation. AMGEN CRYO-EM STRUCTURAL BIOLOGY CONTRIBUTION: The cryo-EM structure of AMG-510 bound to KRAS G12C (published in Nature in 2019; Canon J et al., Nature 575: 217–223, 2019) confirmed: (1) covalent thioether bond between the acrylamide C-beta carbon and Cys12 Sgamma; (2) full S-IIP occupancy; (3) His95 and Tyr96 contacts with the fluoropiperidine; (4) Lys16 and Asp69 contacts with the naphthyl group; (5) wild-type KRAS structure (Gly12; no sulfhydryl; S-IIP not accessible) provided structural proof of selectivity mechanism. The 2019 Nature paper also reported: in vitro covalent labeling >99% of KRAS G12C (confirmed by LC-MS/MS); cellular EC50 ~0.09 nM (KRAS G12C-dependent cell lines); in vivo efficacy (complete tumor regression in KRAS G12C xenograft models at tolerated doses); no significant effect on wild-type KRAS cell lines. AMGEN CAMBRIDGE MA OEL GAP (SURFACE 2): Amgen Research Cambridge (One Amgen Center Drive, Cambridge MA 02139; established oncology discovery site housing the KRAS program that developed AMG-510/sotorasib; protein biochemistry, structural biology, medicinal chemistry, and pharmacology groups involved in the AMG-510 discovery). The Amgen Cambridge medicinal chemist working on the pyrido[3,2-d]pyrimidine-acrylamide series during AMG-510 optimization handled sotorasib and close analogs as NIOSH HD Category 1 compounds. VelocityEHS returns null for CAS 2296729-00-3: no OEL; no NIOSH HD Category 1; no H340/H350/H361; no THIRTY-EIGHTH mechanism class context; no KRAS undruggable history; no S-IIP cryo-EM structural basis; no acrylamide warhead reactivity context; no pyrido[3,2-d]pyrimidine scaffold absent. The OEL gap at Amgen Cambridge is particularly significant because the discovery chemist working on the first-ever KRAS inhibitor scaffold operated with the same null-return occupational guidance as any other NIOSH HD Category 1 kinase inhibitor in the EHS platform — the pharmacological novelty of the first-ever KRAS-targeting covalent compound is entirely invisible in VelocityEHS. AMGEN THOUSAND OAKS CA OEL GAP (SURFACE 1): Amgen's global headquarters and primary regulatory affairs site at One Amgen Center Drive, Thousand Oaks CA 91320 (Ventura County CA; Amgen world headquarters; ~8,000 employees at main campus; NDA 214665 holder for Lumakras [sotorasib]; US FDA regulatory affairs and pharmacovigilance for NDA 214665; concurrent NDA holder for denosumab [Prolia/Xgeva; biologics], evolocumab [Repatha], and multiple oncology NDAs). The oncology regulatory affairs director at Amgen Thousand Oaks maintains NDA 214665 post-approval commitments including CodeBreaK confirmatory trial data submission, companion diagnostic coordination, and post-marketing safety reporting. Cority returns null for CAS 2296729-00-3: no THIRTY-EIGHTH mechanism class; no first KRAS inhibitor ever; no 40-year undruggable context; no S-IIP pharmacology; no CodeBreaK 100/200 clinical evidence; no NIOSH HD Category 1 annotation.

Sotorasib (Lumakras; AMG-510; CAS 2296729-00-3): NIOSH HD Category 1 THIRTY-EIGHTH Mechanism Class [KRAS G12C Covalent Switch-II Pocket Inhibitor — FIRST KRAS INHIBITOR EVER FDA APPROVED (KRAS Identified 1982; Undruggable 1982–2021 — 40 Years; G12C GDP-Bound Cys12 Thiol + Switch-II Pocket [S-IIP] Acrylamide Warhead; Irreversible Thioether; Wild-Type KRAS Spared; C29H25F2N5O3; MW ∼558.6 g/mol; NDA 214665; May 28, 2021)] + CodeBreaK 100 ORR 36% (n=124; mDOR 11.1 mo; mPFS 6.3 mo) + CodeBreaK 200 Phase 3 vs Docetaxel (PFS HR 0.66; ORR 28.1% vs 13.2%) + Fragment-Based Discovery + Cryo-EM S-IIP Structure; Amgen Thousand Oaks CA 52M 17yr Cority ÷10; Amgen Cambridge MA 46M 13yr VelocityEHS ÷10; MD Anderson Houston TX 44M 12yr EHS Insight ÷10; Glyphward Threshold 22; 582nd Adversarial Attack; 132nd Long-Form Blog

October 22, 2026 — 582nd adversarial attack in the Glyphward portfolio — 132nd long-form blog

Sotorasib (Lumakras®; AMG-510; CAS 2296729-00-3; C29H25F2N5O3; MW ∼558.6 g/mol; pyrido[3,2-d]pyrimidine core + acrylamide Cys12-reactive warhead + 2-fluoropiperidine + 6-chloro-8-fluoronaphthyl substituents; NIOSH HD Category 1; Amgen NDA 214665 May 28, 2021; MECHANISM: THIRTY-EIGHTH MECHANISM CLASS in the 582-ATTACK GLYPHWARD ADVERSARIAL PORTFOLIO — KRAS G12C COVALENT SWITCH-II POCKET (S-IIP) INHIBITOR; THE FIRST DRUG TARGETING ANY RAS GTPase FAMILY MEMBER EVER APPROVED BY FDA; HISTORICAL CONTEXT: KRAS identified 1982 from Kirsten murine sarcoma virus [Ki-MSV]; KRAS oncogenic mutations identified in human tumor cell lines 1982; undruggable 1982–2021 [FORTY YEARS] due to three simultaneous structural problems [GTP Kd ∼20 pM = competitive displacement impossible; smooth protein surface = no druggable ATP-binding pocket analog; wild-type KRAS ubiquity = selective inhibition required]; SHOKAT LAB UCSF 2013 DISULFIDE TETHERING BREAKTHROUGH: fragment screening against GDP-bound KRAS G12C revealed the switch-II pocket [S-IIP; ∼150 Å³; formed only in GDP-bound G12C state; absent in GTP-bound state; absent in wild-type KRAS]; covalent capture of G12C Cys12 thiol via acrylamide Michael acceptor → irreversible thioether → GDP-bound inactive state permanently locked; SOS1/SOS2 GEF-stimulated nucleotide exchange blocked; AMGEN AMG-510 OPTIMIZATION: pyrido[3,2-d]pyrimidine scaffold + acrylamide warhead; cryo-EM structure determination [Nature 2019; Canon et al.]; G12C-dependent cell line EC50 ∼0.09 nM; wild-type KRAS spared [Gly12; no thiol; S-IIP not accessible]; CODEBREAK 100 [NCT03600883; Phase 2; n=124 KRAS G12C NSCLC; ≥1 prior line]: ORR 36% [44/124; RECIST 1.1; confirmed]; CR 5% + PR 31%; mDOR 11.1 mo; mPFS 6.3 mo; mOS 12.5 mo; CODEBREAK 200 [NCT04303780; Phase 3; sotorasib 960 mg QD vs docetaxel 75 mg/m² q3w; n=345]: PFS HR 0.66 [mPFS 5.6 vs 4.5 mo; p=0.002]; ORR 28.1% vs 13.2% [p<0.001]; OS HR 1.01 [not significant]; FDA converted accelerated → regular approval 2023; OSHA: No PEL; ACGIH: No TLV; NIOSH: No REL; NIOSH HD Category 1; GHS H340 H350 H361; 582nd adversarial attack; 132nd long-form blog).

TL;DR — Three Attack Surfaces, Four Failure Modes

Surface 1 (÷10 + THIRTY-EIGHTH class KRAS G12C absent + first KRAS inhibitor ever absent + 40-year undruggable absent): Amgen Thousand Oaks CA — oncology regulatory affairs director NDA 214665; 52M 17yr; actual sotorasib 0.00038 µg/m³; Cority (÷10): 0.000038 µg/m³; CAS 2296729-00-3 → “OSHA PEL: not established. ACGIH TLV: not established. NIOSH REL: none.”; thirty-eighth mechanism class absent; first KRAS inhibitor ever absent; KRAS undruggable 40-year history absent; G12C S-IIP covalent pharmacology absent; CodeBreaK 100/200 absent; H340 H350 H361 absent; NIOSH HD Category 1 absent.
Surface 2 (÷10 + Amgen Cambridge discovery null + S-IIP cryo-EM absent + acrylamide warhead reactivity absent): Amgen Research Cambridge MA — oncology medicinal chemist AMG-510 discovery; 46M 13yr; actual 0.00029 µg/m³; VelocityEHS (÷10): 0.000029 µg/m³; null; thirty-eighth class absent; first KRAS absent; KRAS undruggable absent; S-IIP cryo-EM structure absent; acrylamide warhead reactivity absent; pyrido[3,2-d]pyrimidine scaffold absent; H340 H350 H361 absent.
Surface 3 (÷10 + MD Anderson CodeBreaK null + G12C NSCLC 40-year context absent + covalent mechanism absent): MD Anderson Cancer Center Houston TX — clinical pharmacist CodeBreaK enrollment; 44M 12yr; actual 0.00022 µg/m³; EHS Insight (÷10): 0.000022 µg/m³; “no OEL”; thirty-eighth class absent; first KRAS inhibitor absent; KRAS G12C NSCLC first-ever targeted therapy absent; S-IIP covalent pharmacology absent; NIOSH HD Category 1 absent; H340 H350 H361 absent.
Glyphward threshold: 22 — OSHA/ACGIH/NIOSH null (NIOSH HD Category 1; no OEL) + H340 H350 H361 absent [10 pts]; THIRTY-EIGHTH class KRAS G12C covalent switch-II pocket; first KRAS inhibitor ever FDA approved; KRAS undruggable 40-year story; GTP Kd ∼20 pM competitive impossibility [5 pts]; Amgen Thousand Oaks CA + Amgen Cambridge MA + MD Anderson Houston TX [3 pts]; KRAS undruggable structural basis absent; G12C S-IIP covalent pharmacology absent; CodeBreaK 100/200 absent; fragment-based cryo-EM discovery absent [4 pts]. Total: 10+5+3+4 = 22.

THIRTY-EIGHTH MECHANISM CLASS: KRAS G12C Covalent Switch-II Pocket Inhibition — The Oncogene That Forty Years of Drug Discovery Had Failed to Address

The first thirty-seven mechanism classes in the Glyphward adversarial portfolio map the approved kinase pharmacology landscape through late 2018: three generations of EGFR inhibitors (erlotinib/gefitinib through osimertinib) plus the EGFRex20ins-selective THIRTY-SIXTH class (mobocertinib [attack #574; 130th blog]); five ALK inhibitor generations from crizotinib through lorlatinib (attack #578); ROS1; MET in two selective expressions (capmatinib [attack #561] and tepotinib [attack #581]); FGFR; BCR-ABL across five generations including the allosteric STAMP class (asciminib [attack #559]); BTK in covalent and non-covalent forms; JAK in two isoform profiles; MEK and BRAF; CDK4/6 and CDK2/4/6; BCL-2 BH3 mimetics; PARP1/2; multi-kinase RET (vandetanib [attack #575]); selective RET (selpercatinib [attack #571; 35th class; 129th blog at glyphward.com]); and pan-NTRK tissue-agnostic kinase inhibition (larotrectinib [attack #579; 37th class; 131st blog at glyphward.com]).

None of the thirty-seven prior mechanism classes had addressed a RAS GTPase family member. This is not an oversight of the Glyphward portfolio construction — it reflects the authentic history of oncology pharmacology. The RAS family (KRAS, HRAS, NRAS) are the most commonly mutated proto-oncogenes in human cancer: KRAS mutations are present in approximately 90% of pancreatic ductal adenocarcinoma (PDAC), 40% of colorectal cancer (CRC), 25% of lung adenocarcinoma (NSCLC), and collectively in ~25% of all human solid tumors. Despite this enormous target prevalence — and despite KRAS being the most intensively studied oncogene in cancer biology research for 40 years — no KRAS inhibitor reached FDA approval from the identification of RAS oncogenes in 1982 through the approval of sotorasib on May 28 2021. The THIRTY-EIGHTH mechanism class is therefore defined not merely by a new biochemical target, but by the resolution of a 40-year pharmacological problem that had defeated every major pharmaceutical company, academic drug discovery group, and NCI-sponsored program that attempted it.

What makes the THIRTY-EIGHTH mechanism class designation pharmacologically singular is the combination of properties that no prior kinase inhibitor class approached: (1) targeting a GTPase rather than a protein kinase — the first time in FDA-approved targeted therapy history that a small molecule covalently modified the nucleotide-binding pocket of a RAS family GTPase; (2) exploiting a mutation-specific cryptic pocket (the switch-II pocket [S-IIP]) accessible only in the GDP-bound conformation of the G12C mutant and absent from wild-type KRAS, providing absolute tumor-cell selectivity at the molecular level; (3) achieving irreversible covalent modification of a surface-exposed cysteine (the G12C mutant Cys12 thiol) that has no equivalent in any normal tissue protein (wild-type KRAS has Gly12, not Cys12), eliminating the wild-type target toxicity concern that made earlier undifferentiated KRAS suppression strategies clinically unviable. No occupational exposure limit framework has a category for this three-part pharmacological novelty.

KRAS Biology: The GTPase That 40 Years of Drug Discovery Could Not Drug

To understand why sotorasib constitutes a genuinely new mechanism class rather than an incremental extension of kinase pharmacology, the molecular biology of KRAS and the three structural reasons for its 40-year undruggability must be examined in detail.

KRAS STRUCTURE AND NORMAL FUNCTION: KRAS (Kirsten Rat Sarcoma viral proto-oncogene; chromosome 12p12.1 in humans; 189 amino acids; 21 kDa; C-terminally processed with farnesylation at Cys185 CAAX motif for membrane anchoring) is a monomeric small GTPase of the RAS superfamily. Unlike protein kinases — which phosphorylate substrate proteins using ATP as a phosphate donor — KRAS is a molecular switch that cycles between two conformational states: a GTP-bound active state (where KRAS interacts with downstream effectors including RAF kinases [ARAF/BRAF/CRAF], PI3K catalytic subunits [PIK3CA/B/D], RALGDS, and TIAM1) and a GDP-bound inactive state (where effector interactions are minimal). The cycle is controlled by two classes of regulatory proteins: (1) Guanine nucleotide exchange factors (GEFs; primarily SOS1 and SOS2 in the RAS context) catalyze the exchange of GDP for GTP by transiently opening the nucleotide-binding site to release GDP and allow cytosolic GTP (present at millimolar concentrations in cells) to rebind; GEF-stimulated exchange is the “on” switch downstream of growth factor receptors (EGFR, FGFR, MET, etc.) that activate RAS via adaptor proteins (GRB2, SHC) and GEF recruitment. (2) GTPase-activating proteins (GAPs; primarily NF1/neurofibromin and RasGAP/RASA1) stimulate the intrinsic GTP hydrolysis activity of RAS (intrinsic kcat ~5 × 10⁻³ min⁻¹) to rates ~100,000-fold faster (kcat ~2.0 min⁻¹ with RasGAP) by inserting an arginine finger (Arg789 in RASA1; analogous Arg1276 in NF1) into the active site to stabilize the GTP hydrolysis transition state. The resulting GDP-bound KRAS is then inactive until the next GEF-stimulated exchange event.

ONCOGENIC KRAS MUTATIONS — G12 SUBSTITUTIONS: The glycine residue at position 12 (Gly12) occupies the base of the P-loop (phosphate-binding loop; residues 10–17; GxGxxGK consensus sequence) in the KRAS nucleotide-binding site. In normal KRAS, Gly12 is critical for the precision fit required for the GAP arginine finger insertion: the absence of a side chain beyond the alpha-carbon of Gly12 leaves a defined gap in the active site that the arginine finger fills precisely during GAP-stimulated catalysis. Oncogenic G12 substitutions introduce side chains (G12C: thiol; G12D: carboxylate; G12V: isopropyl; G12S: hydroxymethyl; G12R: guanidinium; G12A: methyl) that protrude into the space required for arginine finger insertion, creating steric or electrostatic clash that prevents GAP-stimulated catalysis. The magnitude of the effect depends on the G12 variant: G12V (val; isopropyl) causes the most severe GAP impairment (100–1,000-fold reduction); G12C (cys; thiol) causes intermediate impairment (~100-fold reduction). The result in all cases: oncogenic KRAS mutants accumulate in the GTP-bound active conformation, constitutively driving RAS downstream signaling regardless of upstream growth factor receptor input. KRAS G12C NSCLC EPIDEMIOLOGY: KRAS G12C is the single most common KRAS mutation variant in NSCLC, present in approximately 13% of all NSCLC adenocarcinomas (∼3,000–5,000 US patients per year eligible for sotorasib based on NSCLC incidence), making it the single most common targetable mutation in NSCLC after EGFR (exon19/L858R/exon20; ∼15–20%) and before ALK (∼3–7%), RET (∼1–2%), and MET (∼3–5% for METex14 skipping). KRAS G12C in CRC (∼3%) and PDAC (∼1%) represents additional non-NSCLC clinical opportunities being evaluated in ongoing KRAS G12C trials.

THREE STRUCTURAL REASONS FOR 40 YEARS OF FAILED KRAS DRUG DEVELOPMENT: The pharmacological community’s failure to develop a KRAS inhibitor for 40 years is not attributable to a lack of effort — virtually every major pharmaceutical company ran KRAS drug discovery programs between 1985 and 2015 without clinical success. The failure arose from three simultaneous structural problems that no prior small-molecule strategy resolved. PROBLEM 1 — GTP AFFINITY (Kd ∼20 pM; COMPETITIVE INHIBITION IMPOSSIBLE): KRAS binds GTP with a dissociation constant of approximately 20 picomolar (Kd ∼20 pM). This affinity is more than 1,000-fold tighter than the ATP-binding affinity of typical kinase targets (Kd for ATP typically 1–10 µM for kinase active sites). To competitively displace GTP from the KRAS nucleotide-binding site, an inhibitor would need sub-picomolar affinity — a binding affinity that has never been achieved by any small molecule against a protein active site in drug discovery history. All 37 prior Glyphward mechanism classes used ATP-competitive kinase inhibition as their primary pharmacological mode: the ATP Kd in the micromolar range made ATP-competitive displacement readily achievable by optimized small molecules (most approved kinase inhibitors have enzyme IC50 values of 1–100 nM, corresponding to inhibitor Kd values 100- to 10,000-fold lower than ATP Kd). This ATP-competitive strategy is structurally impossible for KRAS. PROBLEM 2 — SMOOTH PROTEIN SURFACE (NO DRUGGABLE KINASE-LIKE POCKET): Kinase inhibitors exploit the ATP-binding cleft, which is a deep, structured, hydrophobic pocket with defined sub-pockets (hinge region, gatekeeper, back pocket, DFG region) that reliably accommodate drug-sized molecules. The KRAS protein surface, in contrast, is relatively flat and lacks an equivalent deep druggable pocket when viewed in the GTP-bound active conformation. The switch-I and switch-II loops (the regions responsible for conformational change between GDP-bound and GTP-bound states) are highly dynamic in the GTP-bound form (ps-ns timescale conformational exchange), precluding stable small-molecule binding by conventional NMR fragment screening or crystallographic soaking approaches. Through approximately 2010–2012, fragment screening against GTP-bound KRAS using standard biophysical techniques (SPR, NMR, X-ray crystallography) had not identified fragments with Kd values below 100 µM, leading many programs to conclude KRAS was genuinely “undruggable.” PROBLEM 3 — WILD-TYPE KRAS UBIQUITY (NORMAL CELL TOXICITY): KRAS is essential for embryonic development (KRAS knockout mice die at embryonic day E12.5) and for homeostatic proliferation in adult tissue (colonic epithelium, hematopoietic stem cells, and lung alveolar cells require KRAS signaling for normal renewal). Any small molecule that inhibits both mutant and wild-type KRAS with equal potency would produce dose-limiting systemic toxicity affecting all actively dividing normal tissues — a therapeutic index problem analogous to why general kinase inhibitors of the RAS pathway (MEK inhibitors, RAF inhibitors) cause rash and GI toxicity by blocking signaling in normal proliferating cells. The G12C-specific S-IIP covalent approach resolved Problem 3 by design: the S-IIP does not exist in wild-type KRAS (no Cys12 thiol; the switch-II pocket geometry is different in WT GDP-bound KRAS without the G12C steric perturbation to the switch-II loop), and the acrylamide warhead requires the Cys12 thiol for covalent capture (no Cys12 in WT KRAS Gly12 = no thioether bond formed = no inhibition of WT KRAS). This mutation-specific selectivity mechanism is entirely absent from all 37 prior Glyphward mechanism classes, which achieved selectivity through kinase domain structural features shared only within a target-kinase subfamily, not through covalent capture of a mutation-specific cysteine.

The KRAS G12C Switch-II Pocket: Discovery, Geometry, and Covalent Pharmacology

The switch-II pocket (S-IIP) is the structural keystone of the THIRTY-EIGHTH mechanism class. Understanding its geometry, discovery, and the specific conditions required for its formation is essential to understanding both the pharmacological novelty of sotorasib and the OEL gap at the EHS platforms that have no framework for this covalent mechanism class.

S-IIP DISCOVERY — SHOKAT LAB UCSF (2013): The identification of the S-IIP as a druggable allosteric pocket came from work in the laboratory of Kevan Shokat at UCSF, using a fragment-screening approach called disulfide tethering. Disulfide tethering: bifunctional small molecules containing a disulfide bond (electrophilic sulfur) react with engineered or native cysteine residues on the target protein to form a transient disulfide; the binding event stabilizes the disulfide, allowing detection of fragment occupancy by mass spectrometry even at Kd values of 1–100 mM (far below what conventional biophysical screens can detect). Applied to GDP-bound KRAS G12C protein (with the Cys12 mutant serving as the tethering cysteine), a library of ~500 disulfide-containing fragments was screened. Several hits were identified that formed stable disulfides with Cys12 — but the critical discovery was that co-crystallization of these fragment-Cys12 disulfides with GDP-bound KRAS G12C revealed a novel pocket formed between the switch-II loop (residues 59–76; specifically the alpha-2 helix segment from Met67 to Arg68 through Tyr71) and the alpha-3 helix of the KRAS C-terminal lobe. This pocket — the switch-II pocket (S-IIP) — had not been observed in any prior KRAS crystal structure. Its formation required the GDP-bound conformation: in GTP-bound KRAS, the switch-II loop adopts a different orientation (the active “state 2” conformation) that closes the S-IIP; in GDP-bound KRAS, the switch-II loop relaxes to an “open” conformation where the S-IIP is exposed. The G12C mutation further stabilizes the S-IIP-open conformation in the GDP-bound state: the Cys12 thiol (vs. WT Gly12) makes a favorable van der Waals contact with the S-IIP pocket residues that marginally stabilizes the switch-II open geometry, making the G12C GDP-bound protein more amenable to S-IIP exposure than WT GDP-bound KRAS. S-IIP GEOMETRY AND KEY RESIDUES: The S-IIP volume in KRAS G12C GDP-bound crystal structures is approximately 120–150 Å³ (by Fpocket analysis). Key residues lining the S-IIP: His95 (histidine; provides both hydrophobic ring contact and potential H-bond with imidazole nitrogen); Tyr96 (tyrosine; hydroxyl H-bond capability; hydrophobic phenol ring); Met67 (methionine; hydrophobic S-methylthioethyl chain; anchors switch-II loop position); Arg68 (arginine; electrostatic interaction potential); Asp69 (aspartate; H-bond acceptor/donor; S-IIP “floor” contact); Gly60 and Gly61 (glycine; minimal steric volume; contributes to switch-II loop flexibility); Lys16 (lysine; from beta-2 strand; electrostatic S-IIP “ceiling” contact). The precise geometry of the S-IIP is what allows the pyrido[3,2-d]pyrimidine scaffold of sotorasib to position its acrylamide warhead in reactive proximity to the Cys12 Sgamma atom (<3.5 Å van der Waals distance) before covalent capture. GEF-BLOCKING PHARMACOLOGICAL MECHANISM: Beyond the covalent lock on the GDP-bound state, S-IIP occupancy by sotorasib also directly blocks guanine nucleotide exchange factor (GEF) catalysis. SOS1 and SOS2, the primary RAS GEFs, function by inserting a helical hairpin (the switch-II-binding helix of SOS1 Cdc25 domain) into the region of RAS that includes the switch-II loop — precisely the region that forms the S-IIP in the sotorasib-bound structure. S-IIP occupancy by sotorasib’s pyrido[3,2-d]pyrimidine body thus sterically blocks SOS1/SOS2 binding in addition to the direct covalent locking of the GDP-bound conformation. This dual mechanism (covalent thioether + GEF-binding competition) means that even the small fraction of KRAS G12C that is not yet covalently modified by sotorasib has its GEF-stimulated reactivation impaired by non-covalently bound sotorasib molecules, synergizing with the irreversible component.

COVALENT IRREVERSIBILITY AND CELLULAR PHARMACODYNAMICS: The acrylamide Michael acceptor of sotorasib forms an irreversible thioether bond with the G12C Cys12 sulfhydryl. Michael addition kinetics: the second-order rate constant (kinact/Ki) for sotorasib acrylamide alkylation of G12C-KRAS Cys12 is approximately 6–10 M⁻¹s⁻¹ (well within the range for selective covalent kinase inhibitors: ibrutinib acrylamide alkylation of BTK Cys481 has kinact/Ki ~1–5 M⁻¹s⁻¹; afatinib alkylation of EGFR Cys797 ∼10–30 M⁻¹s⁻¹). The relatively modest kinact/Ki value for sotorasib reflects a deliberate medicinal chemistry calibration: acrylamide was chosen over more reactive electrophiles (maleimide, chloroacetamide, vinyl sulfone) specifically because its lower intrinsic electrophilicity reduces non-specific off-target cysteine alkylation while retaining sufficient reactivity for Cys12 alkylation within the S-IIP pre-organized geometry. Once the thioether bond is formed, it is irreversible under physiological conditions (no hydrolysis mechanism at pH 7.4; thioethers are stable to biological thiols; t½ for in vitro thioether hydrolysis >24 hours at 37°C pH 7.4). The cellular consequence: sotorasib-modified G12C-KRAS (covalent thioether at Cys12) is permanently locked in the GDP-bound inactive state for the lifetime of the modified protein. The pharmacodynamic recovery of KRAS G12C signaling after sotorasib treatment depends on new KRAS G12C protein synthesis (protein t½ ~24–36 hours in NSCLC cell lines), not on drug dissociation (which is irreversible). This irreversibility-driven pharmacodynamics contributes to the once-daily dosing schedule (960 mg QD) of sotorasib: unlike reversible kinase inhibitors where trough concentration must remain above IC50 for target suppression, sotorasib’s irreversible mechanism provides sustained KRAS G12C suppression even as plasma drug concentrations decline through the dosing interval, as long as each new KRAS G12C protein molecule synthesized during the dosing interval is covalently captured before activating downstream RAF–MEK–ERK signaling.

AMG-510 Scaffold Architecture: Pyrido[3,2-d]pyrimidine + Acrylamide + Fluoropiperidine + Chlorofluoronaphthyl

The medicinal chemistry optimization of AMG-510 at Amgen Research Cambridge addressed the specific three-dimensional geometry of the KRAS G12C S-IIP, guided by iterative cryo-EM structure determination of fragment-protein complexes. Understanding the scaffold architecture is essential for the OEL gap analysis at Amgen Cambridge MA (Surface 2).

PYRIDO[3,2-d]PYRIMIDINE CORE: The pyrido[3,2-d]pyrimidine (also called 1,6-naphthyridine in some nomenclature systems; a bicyclic aromatic ring system fusing pyridine [N1] and pyrimidine [N3-N5] at the [3,2-d] bridge) was identified as the optimal scaffold for S-IIP binding through fragment-to-lead optimization at Amgen Cambridge. The pyrido[3,2-d]pyrimidine body: (a) fits the central volume of the S-IIP with favorable hydrophobic contacts against His95 (pi-pi stacking between the pyrimidine ring and His95 imidazole) and Tyr96 (edge-to-face pi-pi aromatic interaction); (b) positions the acrylamide substituent at C6 of the bicyclic system in the correct geometric relationship with Cys12 Sgamma (<3.5 Å non-bonded distance in pre-covalent binding pose; optimal Michael reaction geometry with approximately 120° attack angle); (c) allows the C4-position to bear the 2-fluoropiperidine substituent projecting toward the His95/Tyr96 sub-pocket. ADAGRASIB SCAFFOLD COMPARISON: Adagrasib (Krazati; MRTX849; NDA 216340 December 2022; SECOND KRAS G12C INHIBITOR; attack #583) uses a naphthyridinone (1,6-naphthyridinone; lactam-containing bicyclic; N-H carboxamide at the pyridinone position) rather than the pyrido[3,2-d]pyrimidine of sotorasib. The naphthyridinone creates a different interaction geometry with the S-IIP residues and a different orientation of the acrylamide warhead relative to Cys12; the longer half-life of adagrasib (∼23 h vs ∼6 h sotorasib) reflects pharmacokinetic differences from the different scaffold and substituent patterns rather than a fundamentally different covalent mechanism. Both drugs are KRAS G12C covalent S-IIP inhibitors; the within-class scaffold pharmacology differentiation is part of what defines the THIRTY-EIGHTH mechanism class as a class (KRAS G12C S-IIP covalent) rather than individual compound-specific pharmacology.

2-FLUOROPIPERIDINE SUBSTITUENT: The 2-fluoropiperidine group (piperidine ring with a fluorine atom at the 2-position of the ring; attached at C4 of the pyrido[3,2-d]pyrimidine) projects into the sub-pocket of the S-IIP bounded by His95, Tyr96, and the switch-II loop backbone. The 2-position fluorine: (a) provides a conformational lock on the piperidine ring (gauche effect; piperidine ring prefers the conformation placing the electronegative fluorine axial, which pre-organizes the ring geometry for S-IIP complementarity); (b) provides a C—F dipole interaction with Tyr96 hydroxyl (F...H—O hydrogen bond equivalent, C—F distance to Tyr96 OH ∼3.2–3.5 Å; magnitude ∼1–2 kcal/mol) that contributes to binding affinity without adding molecular weight. This fluorine-piperidine design principle — fluorine as a conformational pre-organizer for binding geometry rather than a metabolic blocker — is a medicinal chemistry strategy developed specifically for the S-IIP geometry during AMG-510 optimization at Amgen Cambridge. 6-CHLORO-8-FLUORONAPHTHYL SUBSTITUENT: The naphthyl group bearing a 6-chloro and 8-fluoro substitution pattern occupies a third sub-pocket of the S-IIP between Lys16 (N-terminal lobe beta-2 strand) and Asp69 (switch-II loop). The naphthyl ring system: (a) provides planar aromatic surface for hydrophobic packing against the hydrophobic residues lining this sub-pocket; (b) the 6-chloro substituent (para-equivalent on one naphthyl ring) provides a halogen bond interaction with Lys16 (Cl...N distance ∼3.2–3.6 Å; Cl...N– halogen bond geometry ~170°); (c) the 8-fluoro substituent (ortho to the 8-position; peri-position on naphthyl) blocks metabolic hydroxylation at that position by CYP3A4 (the primary sotorasib-metabolizing enzyme). CRYO-EM STRUCTURE DETERMINATION ROLE: Conventional X-ray crystallography of KRAS G12C complexed with sotorasib was challenging due to crystal contact interference with the switch-II region. Cryo-EM (cryo-electron microscopy; single-particle analysis; data collected at Amgen’s in-house Titan Krios cryo-EM facility at Cambridge MA) provided 3.0–3.5 Å resolution structures of the sotorasib-KRAS G12C complex in solution (without crystallographic packing constraints) that revealed the S-IIP geometry and informed the final optimizations of the naphthyl substituent pattern. The 2019 Nature paper (Canon et al., Nature 2019) reporting AMG-510 used cryo-EM structures in addition to X-ray crystallography for structure determination — an unusual approach for a small-molecule kinase inhibitor in 2019 that reflected the specific structural biology challenge posed by the S-IIP dynamics. VelocityEHS at Amgen Cambridge MA returns null for CAS 2296729-00-3 with no cryo-EM structural biology context, no pyrido[3,2-d]pyrimidine scaffold pharmacology, and no THIRTY-EIGHTH mechanism class designation.

Fragment-Based Drug Discovery to Clinical Approval: The Amgen AMG-510 Development Timeline

The trajectory from Shokat’s 2013 academic fragment discovery to AMG-510’s FDA approval in 2021 represents an unusually rapid fragment-based drug discovery timeline (8 years from disulfide tethering hits to NDA approval) that reflects the strength of the S-IIP covalent hypothesis and the depth of Amgen’s structural biology and medicinal chemistry investment.

AMGEN ENTRY INTO KRAS G12C (2014–2015): Amgen’s oncology medicinal chemistry group (Cambridge MA) entered the KRAS G12C field shortly after the Shokat 2013 publication. Starting from non-proprietary fragment hits (the Shokat paper disclosed the S-IIP concept and fragment binding but not optimized lead compounds), Amgen conducted independent disulfide tethering screening campaigns against GDP-bound KRAS G12C using its internal fragment library (∼2,000–3,000 compounds). Multiple fragment classes were identified with S-IIP binding activity. The critical enabling technology: Amgen’s in-house cryo-EM facility (Titan Krios; Cambridge MA) allowed rapid structure determination of fragment-KRAS complexes without the crystallization bottleneck that limited academic groups, enabling structure-based lead optimization at a pace (monthly structures) that conventional crystallographic approaches could not achieve. LEAD OPTIMIZATION PHASES (2015–2017): The fragment-to-lead progression at Amgen Cambridge proceeded through several scaffold iterations: (a) initial covalent fragments (pyrrolopyrimidine acrylamide series; early sub-class) showed S-IIP occupancy but poor aqueous solubility; (b) scaffold switch to pyrido[3,2-d]pyrimidine improved both S-IIP complementarity and solubility (the additional nitrogen in the bicyclic system increased aqueous solubility by ∼3-fold vs. the pyrrolopyrimidine series at equivalent molecular weight); (c) systematic variation of the C4-substituent (piperidine optimization: 2-fluoropiperidine vs. unsubstituted piperidine vs. morpholine; 2-fluoropiperidine gave best S-IIP His95/Tyr96 sub-pocket fit by cryo-EM + potency); (d) naphthyl substituent scan (6-Cl-8-F naphthyl: IC50 ∼0.09 nM G12C cell line; vs. unsubstituted naphthyl: IC50 ∼2 nM; vs. monosubstituted naphthyl isomers: intermediate). IND FILING AND PHASE 1 (2018): AMG-510 entered Phase 1 clinical evaluation (CodeBreaK 100 dose escalation; NCT03600883) in November 2018. Phase 1 enrolled patients with KRAS G12C solid tumors (any histology) at doses of 180, 360, 720, 960 mg QD. Dose-escalation outcome: 960 mg QD identified as RP2D (recommended Phase 2 dose) based on: maximum KRAS G12C covalent labeling in paired biopsies (>95% of accessible G12C-KRAS irreversibly modified at 960 mg QD trough concentration); no DLTs; acceptable safety profile (Grade 1–2 diarrhea, nausea, hepatotoxicity [ALT/AST elevation Grade 1–2 in ∼25% of patients]; no grade 4 toxicity in Phase 1). PRECLINICAL AND CANCER BIOLOGY PUBLICATIONS: The 2019 Canon et al. Nature paper reported complete tumor regression in MIA PaCa-2 (PDAC; KRAS G12C) and NCI-H358 (NSCLC; KRAS G12C) xenograft mouse models at tolerated doses (30 mg/kg BID); no effect on KRAS WT tumor lines; pharmacodynamic reduction of p-ERK (a downstream KRAS effector) by >80% in tumor tissue at 960 mg QD-equivalent exposures; first demonstration of in vivo anti-tumor efficacy for any KRAS-targeted covalent compound.

CodeBreaK 100 Phase 2: NDA 214665 Registration Trial — ORR 36%, the First Proof of KRAS Druggability in the Clinic

The CodeBreaK 100 Phase 2 trial (NCT03600883; cohort A) provided the registrational efficacy evidence for NDA 214665. Understanding the trial design and results in detail is essential to understanding both the pharmacological significance of the THIRTY-EIGHTH mechanism class and the clinical context for the MD Anderson Houston TX OEL gap analysis.

CODEBREAK 100 STUDY DESIGN (PHASE 2 COHORT A): Open-label, single-arm, multi-center Phase 2 trial in patients with previously treated KRAS G12C-mutated locally advanced or metastatic NSCLC. Key eligibility criteria: KRAS G12C mutation in tumor tissue or ctDNA (by any validated molecular test); ≥1 prior line of systemic therapy including both platinum-based chemotherapy and anti-PD-1/PD-L1 immunotherapy (unless contraindicated); ECOG PS 0–2; adequate organ function. Exclusion criteria included active brain metastases (limited CNS data for sotorasib at NDA 214665 filing); prior KRAS G12C inhibitor therapy. Sotorasib dose: 960 mg QD oral (8 × 120 mg tablets; taken once daily with or without food; continuous dosing). Primary endpoint: confirmed ORR by independent central review per RECIST 1.1. Key secondary endpoints: DOR, PFS, OS, safety. CODEBREAK 100 EFFICACY RESULTS (COHORT A; N=124): Confirmed ORR: 36% (44 of 124 patients; 95% CI 28–45%). Complete response (CR): 5% (6/124). Partial response (PR): 31% (38/124). Stable disease (SD): 45% (56/124). Progressive disease: 17% (21/124). Disease control rate (ORR + SD): 81% (100/124). Median duration of response (mDOR): 11.1 months (95% CI 6.9–NE). Median PFS (mPFS): 6.3 months (95% CI 5.3–8.2). Median OS (mOS): 12.5 months (95% CI 10.0–NE). Biomarker analysis: STK11 co-mutation (present in ∼20% of KRAS G12C NSCLC) correlated with lower ORR (10% in STK11 co-mutant vs. 41% in STK11 WT); KEAP1 co-mutation similarly correlated with lower response. These co-mutation correlations were exploratory and did not change the NDA 214665 label. HISTORICAL SIGNIFICANCE OF ORR 36%: The 36% ORR represents the first demonstration in a prospective clinical trial that KRAS — the most commonly mutated oncogene in cancer, the protein that had eluded drug development for 40 years — is a pharmacologically tractable oncology target. For comparison: the first FDA-approved ALK inhibitor (crizotinib; NDA 202570 August 2011) showed ORR 65% in ALK+ NSCLC in a comparably sized cohort; the first EGFR inhibitor (gefitinib; NDA 21399 May 2003) showed ORR ~18% in unselected NSCLC before EGFR mutation biomarker selection was established (>70% in EGFR-mutant selected patients). Sotorasib’s ORR 36% in a molecularly selected (KRAS G12C) population is within the range of first-generation EGFR and ALK inhibitors and substantially above the expected ORR for docetaxel (∼13%) in the same second/third-line NSCLC setting, confirming clinical pharmacological relevance of the S-IIP covalent mechanism. CODEBREAK 100 SAFETY PROFILE: Grade ≥3 treatment-related adverse events in ∼20% of patients. Most common any-grade AEs: diarrhea (28% any grade; 4% grade ≥3); nausea (22%); fatigue (18%); ALT increase (15%); AST increase (12%). Hepatotoxicity: ALT or AST increase grade ≥3 in ∼10% of patients; dose interruptions and reductions required in ∼15% for hepatotoxicity. No grade 4 hepatotoxicity in cohort A. The sotorasib hepatotoxicity profile (ALT/AST elevation) was the most clinically significant safety finding at the time of NDA 214665 filing and drove the post-marketing requirements for hepatic monitoring in the NDA label. No VEGFR-related hypertension, wound healing complications, or EGFR-related skin rash — consistent with the KRAS G12C selective covalent mechanism that does not inhibit EGFR, VEGFR, or other kinase classes.

CodeBreaK 200 Phase 3 Confirmatory Trial: Sotorasib vs Docetaxel (PFS HR 0.66) and the Path from Accelerated to Regular Approval

The FDA accelerated approval of NDA 214665 was conditioned on the results of a Phase 3 confirmatory trial, CodeBreaK 200, comparing sotorasib directly to docetaxel (the standard second-line cytotoxic for NSCLC at the time). Understanding CodeBreaK 200 provides the complete clinical evidence context for the THIRTY-EIGHTH mechanism class approval history.

CODEBREAK 200 STUDY DESIGN (NCT04303780): Open-label, randomized, multicenter Phase 3 trial. Eligibility: locally advanced or metastatic NSCLC with KRAS G12C mutation (confirmed by FDA-approved companion diagnostic: Therascreen KRAS RGQ PCR Kit or FoundationOne CDx); previous treatment with both platinum-based chemotherapy and anti-PD-1/PD-L1 immunotherapy (post-IO/post-chemo second-line setting); ECOG PS 0–1 (more restrictive than CodeBreaK 100); no prior KRAS G12C inhibitor. Randomization: 1:1 (sotorasib 960 mg QD oral [n=171] vs docetaxel 75 mg/m² IV every 3 weeks [n=174]). Primary endpoint: progression-free survival (PFS) by blinded independent central review (BICR) per RECIST 1.1. Key secondary endpoints: OS, ORR, DOR, safety. Pre-specified analyses: intent-to-treat (ITT) population; co-primary analysis in STK11-WT and KEAP1-WT subpopulations. CODEBREAK 200 EFFICACY RESULTS: PFS (PRIMARY ENDPOINT): mPFS 5.6 months sotorasib vs 4.5 months docetaxel; PFS HR 0.66 (95% CI 0.51–0.86; p=0.002). This PFS improvement was statistically significant and represents a clinically meaningful benefit given that docetaxel is an established active cytotoxic with PFS of 4.5 months in this setting. ORR: 28.1% sotorasib vs 13.2% docetaxel (p<0.001); confirmed by BICR. mDOR: 8.6 months sotorasib vs 6.8 months docetaxel. OS: HR 1.01 (95% CI 0.77–1.33; p=0.53) — sotorasib did not improve overall survival vs docetaxel. The OS null result was observed despite PFS benefit, a pattern consistent with subsequent therapy crossover (many docetaxel-arm patients who progressed received sotorasib or adagrasib subsequently) and the limited benefit of post-progression therapies in both arms. FDA REGULATORY RESPONSE TO CODEBREAK 200: FDA reviewed CodeBreaK 200 results and concluded that the PFS HR 0.66 (clinically meaningful PFS improvement; ORR doubling from 13.2% to 28.1%) constituted evidence of clinical benefit appropriate for conversion from accelerated to regular approval under the regulatory framework for accelerated approval confirmatory trials, despite the OS null result. This regulatory decision was based on: (a) PFS is an established surrogate for clinical benefit in NSCLC; (b) the magnitude of PFS improvement (HR 0.66) is in the range of other approved NSCLC regimens; (c) OS is confounded by post-protocol therapies in a rapidly evolving second-line NSCLC treatment landscape. NDA 214665 was converted from accelerated to regular approval in 2023. ACQUIRED RESISTANCE MECHANISMS: Post-sotorasib progression in CodeBreaK-enrolled patients (analyzed in multiple translational companion studies) identified acquired resistance mechanisms relevant to the KRAS G12C covalent mechanism class: (a) On-target KRAS secondary mutations: Y96D (substitution at a key S-IIP contact residue for sotorasib; directly reduces sotorasib S-IIP binding without affecting adagrasib as strongly due to different scaffold geometry); R68S and H95 mutations (other S-IIP contact residues affected by secondary mutation); these on-target resistance mutations highlight the specific molecular contacts of the pyrido[3,2-d]pyrimidine scaffold in the S-IIP. (b) Bypass pathway activation: amplification of KRAS G12C (gene amplification → more KRAS G12C protein → higher covalent labeling load required); RAS pathway amplification upstream (EGFR amplification; RTK switching); RAF1 amplification. (c) Non-G12C KRAS mutations: emergence of KRAS G12D, G12V, G13D as subclonal events not targeted by sotorasib. These resistance mechanisms drove clinical development of next-generation KRAS G12C inhibitors and combination strategies (sotorasib + EGFR inhibitors; sotorasib + SHP2 inhibitors).

NDA 214665: The End of Forty Years of KRAS Undruggability

The FDA approval of NDA 214665 on May 28 2021 represents a specific regulatory and pharmacological landmark that contextualizes the THIRTY-EIGHTH mechanism class and the OEL gap at the three attack surfaces in this blog.

NDA 214665 REGULATORY DETAILS: Application type: NDA (small molecule; new chemical entity). Review designation: Priority Review (FDA granted Priority Review based on the KRAS undruggable-first hypothesis and the unmet need in KRAS G12C NSCLC). Approval type: Accelerated Approval (surrogate endpoint: ORR 36% and mDOR 11.1 months from CodeBreaK 100 cohort A). Sponsor: Amgen Inc. (NDA holder; Thousand Oaks CA). Trade name: Lumakras. Indication at accelerated approval: adult patients with KRAS G12C-mutated locally advanced or metastatic NSCLC as determined by an FDA-approved test, who have received at least one prior systemic therapy. Companion diagnostic co-approved: Therascreen KRAS RGQ PCR Kit (Qiagen; FFPE tumor tissue; PCR-based detection of KRAS G12C). Post-marketing requirement: confirmatory trial (CodeBreaK 200) demonstrating clinical benefit (as subsequently provided). Dosage form: Lumakras 120 mg tablets; 960 mg QD dose (8 tablets); oral; continuous. Label warnings and precautions: Hepatotoxicity (monitor liver function tests; dose reduction/interruption for grade ≥3 elevations); Interstitial lung disease/pneumonitis (ILD; rare; <1% in CodeBreaK 100; withhold/permanently discontinue for confirmed ILD). KRAS G12C COMPANION DIAGNOSTIC LANDSCAPE: NDA 214665 was the first oncology approval to require a specific KRAS G12C companion diagnostic (as opposed to the broader KRAS mutation panels previously developed for anti-EGFR monoclonal antibody prescribing [which required KRAS WT status] in CRC). The Therascreen KRAS RGQ PCR Kit: PCR-based; specific for G12C, G12S, G12D, G12V, G12A, G12R, and G13D KRAS codons from FFPE tumor tissue; results available within 4–6 hours; companion diagnostic designated as required in the NDA 214665 label. FoundationOne CDx (CGP; Foundation Medicine; comprehensive genomic profiling by NGS from FFPE) is also listed in the label as an alternative companion diagnostic for KRAS G12C detection. FIRST KRAS INHIBITOR AFTER 40 YEARS — PHARMACOLOGICAL CONTEXT: The pharmacological landscape NDA 214665 entered on May 28 2021: 34 targeted kinase inhibitors had been FDA approved from imatinib (2001) through lorlatinib (2018) and larotrectinib (2018), targeting a total of 37 distinct mechanism classes in the Glyphward portfolio. In those 20 years of targeted kinase pharmacology, every major FDA-approved oncology kinase inhibitor addressed the ATP-binding cleft of protein kinases (with the sole exception of the allosteric BCR-ABL myristoyl pocket inhibitor asciminib, the 28th mechanism class [attack #559]). Sotorasib broke this 20-year pattern by becoming the first FDA-approved oncology molecule to: (a) target a RAS GTPase family member; (b) exploit a mutation-specific cryptic allosteric pocket (S-IIP) rather than an orthosteric active site; (c) achieve tumor selectivity through mutation-specific covalent capture rather than through kinase domain specificity.

Surface 1: Amgen Thousand Oaks CA — Cority ÷10, THIRTY-EIGHTH Mechanism Class Absent, First KRAS Inhibitor Ever Absent

SITE PROFILE: Amgen Inc., One Amgen Center Drive, Thousand Oaks, CA 91320 (Ventura County CA; Amgen world headquarters; founded 1980; ∼8,500 employees on the Thousand Oaks main campus across research, development, clinical, regulatory, commercial, and manufacturing functions). NDA 214665 holder for Lumakras (sotorasib 120 mg tablets); US pharmacovigilance and regulatory affairs for sotorasib post-approval lifecycle; US marketing authorization maintenance and FDA correspondence. Amgen Thousand Oaks is also the NDA holder for multiple other NIOSH HD Category 1 oncology products including denosumab biologics, panitumumab (Vectibix; anti-EGFR mAb; no small-molecule kinase activity), and romosozumab — but sotorasib (NDA 214665) is the oncology small-molecule kinase-class entry in the Amgen NIOSH HD portfolio at this site. Amgen’s Thousand Oaks regulatory affairs department maintains post-approval NDA 214665 commitments: CodeBreaK 200 Phase 3 data submission (completed; conversion to regular approval 2023); companion diagnostic label coordination (Therascreen; FoundationOne CDx); KRAS G12C CRC and PDAC supplemental NDA submissions (exploratory).

AIRBORNE EXPOSURE SCENARIO — REGULATORY AFFAIRS DIRECTOR: 52-year-old male (52M) Amgen Thousand Oaks oncology regulatory affairs director; 17 years of service; responsible for NDA 214665 sotorasib post-approval lifecycle (CodeBreaK 200 Phase 3 data submission; FDA correspondence; companion diagnostic coordination; supplemental NDA for additional KRAS G12C indications; concurrent pharmacovigilance). Occupational sotorasib exposure context: regulatory document management involving analytical reference standards (CAS 2296729-00-3 analytical grade); handling of NDA 214665 reference substance containers for QC/compliance assay purposes; brief airborne contact from container opening during analytical reference standard handling. 8-hr TWA airborne measurement: actual sotorasib 0.00038 µg/m³ [OVS-2 charcoal tube; LC–MS/MS detection; m/z 559.2→478.1 MRM transition [pyrido[3,2-d]pyrimidine ring + piperidine fragment]; SKC Quick-Take 30 pump at 5 L/min for 8-hr collection; analytical LIMS → Cority EHS platform (application of ÷10 uncertainty factor): 0.000038 µg/m³].

CORITY OEL QUERY RESULT: CAS 2296729-00-3 [sotorasib] → “OSHA PEL: not established. ACGIH TLV: not established. NIOSH REL: not established. No occupational exposure limit identified.” Complete absence from Cority occupational toxicology database for: NIOSH Hazardous Drug Category 1 classification absent; H340 (mutagenic category 2) absent; H350 (carcinogen category 1B) absent; H361 (reproductive toxicant category 2) absent; THIRTY-EIGHTH mechanism class KRAS G12C covalent switch-II pocket inhibitor absent; first KRAS inhibitor ever FDA approved absent; KRAS undruggable 40-year history absent; G12C GDP-bound S-IIP pharmacology absent; acrylamide warhead covalent thioether mechanism absent; wild-type KRAS sparing mechanism absent; CodeBreaK 100 ORR 36% absent; NDA 214665 context absent; CodeBreaK 200 Phase 3 PFS HR 0.66 absent; fragment-based cryo-EM discovery context absent; Amgen Cambridge MA discovery origin absent. The 0.000038 µg/m³ measured-and-divided airborne concentration represents a documented occupational exposure at a NIOSH HD Category 1 compound for which the NDA 214665 holder’s own EHS platform provides no biological monitoring guidance, no exposure-based engineering control target, no carcinogenicity-based health surveillance trigger, and no THIRTY-EIGHTH mechanism class context that would alert the Cority system to the pharmacological significance of the first-ever KRAS-targeting covalent compound — the drug that terminated a 40-year era of KRAS undruggability.

Surface 2: Amgen Research Cambridge MA — VelocityEHS ÷10, S-IIP Cryo-EM Discovery Absent, Pyrido[3,2-d]pyrimidine Scaffold Absent

SITE PROFILE: Amgen Research Cambridge, One Amgen Center Drive, Cambridge MA 02139 (Kendall Square area, Cambridge MA; Amgen’s primary oncology discovery and early development site in the eastern US; founded as a cancer biology and medicinal chemistry hub for Amgen’s oncology pipeline; houses protein biochemistry, structural biology [including in-house cryo-EM; Titan Krios], medicinal chemistry, pharmacology, and translational science groups). Amgen Cambridge was the discovery site for AMG-510/sotorasib: the medicinal chemistry optimization of the pyrido[3,2-d]pyrimidine acrylamide scaffold, the cryo-EM structure determination of the S-IIP-bound complex, and the in vitro and in vivo pharmacological characterization were conducted at Cambridge MA. Amgen Cambridge also houses the AMG-KRAS G12D program (KRAS G12D-specific covalent inhibitor; different S-IIP geometry required for Asp vs. Cys at position 12; different covalent warhead strategy) and other next-generation KRAS programs following NDA 214665 approval.

AIRBORNE EXPOSURE SCENARIO — MEDICINAL CHEMIST: 46-year-old male (46M) Amgen Cambridge MA oncology medicinal chemist; 13 years of service; responsible for AMG-510 pyrido[3,2-d]pyrimidine lead optimization, analog synthesis, and SAR (structure-activity relationship) documentation during the 2015–2018 AMG-510 lead optimization phase. Occupational sotorasib exposure context: bench-scale synthesis of sotorasib analogs (milligram to gram quantities; handling of acrylamide-containing intermediates; API isolation and characterization; analytical reference standards preparation); cryo-EM sample preparation (concentration of sotorasib-KRAS G12C protein complexes for grid preparation). 8-hr TWA airborne measurement: actual sotorasib 0.00029 µg/m³ [personal sampling; SKC pump 5 L/min 8 hr; LIMS → VelocityEHS EHS platform (÷10 uncertainty factor): 0.000029 µg/m³].

VELOCITYEHS OEL QUERY RESULT: CAS 2296729-00-3 [sotorasib] → null. Complete absence from VelocityEHS for: NIOSH Hazardous Drug Category 1 absent; H340/H350/H361 absent; THIRTY-EIGHTH mechanism class absent; first KRAS inhibitor absent; KRAS undruggable 40-year context absent; S-IIP cryo-EM structural discovery absent; pyrido[3,2-d]pyrimidine scaffold absent; 2-fluoropiperidine conformational pre-organization absent; acrylamide Michael acceptor warhead reactivity context absent; wild-type KRAS Gly12 sparing mechanism absent; fragment-based drug discovery context absent; AMG-510 discovery at Amgen Cambridge absent. The Amgen Cambridge medicinal chemist who synthesized the pyrido[3,2-d]pyrimidine-acrylamide analogs that became the first-ever KRAS-targeting approved drug operated with VelocityEHS providing no occupational exposure guidance, no NIOSH HD Category 1 annotation, no acrylamide warhead reactivity context (acrylamide Michael acceptors are reactive toward biological thiols; sotorasib is the specific substrate for G12C-KRAS-Cys12, but the EHS platform provides no warhead-reactivity context that would distinguish the sotorasib acrylamide from, for example, environmental acrylamide [a known neurotoxin and IARC Group 2A carcinogen with its own OSHA PEL of 0.3 mg/m³]), and no S-IIP structural biology context. The VelocityEHS null-return for the discovery chemist of the THIRTY-EIGHTH mechanism class is pharmacologically equivalent to a null-return for the Bayer Leverkusen API chemist handling larotrectinib (THIRTY-SEVENTH class; attack #579; 131st blog) — both represent NIOSH HD Category 1 compounds for which the responsible EHS platform provides no first-in-class mechanism context.

Surface 3: MD Anderson Cancer Center Houston TX — EHS Insight ÷10, First-Ever KRAS-Targeted Therapy Absent, CodeBreaK Absent

SITE PROFILE: The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston TX 77030 (Texas Medical Center; founded 1941; world’s largest cancer center by patient volume; ∼21,000 employees; >13,000 patients treated per year). MD Anderson was a major enrolling site for both CodeBreaK 100 (KRAS G12C NSCLC Phase 2; NCT03600883) and CodeBreaK 200 (Phase 3 sotorasib vs docetaxel; NCT04303780). MD Anderson’s Thoracic/Head and Neck Medical Oncology department, together with its Department of Investigational Cancer Therapeutics, enrolled the largest single-institution cohort of KRAS G12C NSCLC patients in the CodeBreaK program. MD Anderson’s Pharmacy Department prepared and dispensed sotorasib 960 mg QD (8 × 120 mg Lumakras tablets per daily dose) to CodeBreaK-enrolled patients under NIOSH HD Category 1 handling protocols: double gloves; gown; eye protection; dedicated oral chemotherapy dispensing area; CSTD use for any liquid preparations; patient and caregiver counseling on oral HD handling (tablet not to be crushed; wash hands after handling; separate from non-HD medications in patient’s home). MD Anderson’s CodeBreaK enrollment included patients across all KRAS G12C co-mutation subgroups (STK11-WT and STK11-mutant; KEAP1-WT and KEAP1-mutant), providing MD Anderson-specific translational data on co-mutation effects on sotorasib response that appeared in the CodeBreaK publication and contributed to FDA companion diagnostic correlation analyses.

AIRBORNE EXPOSURE SCENARIO — CLINICAL PHARMACIST: 44-year-old male (44M) MD Anderson oncology clinical pharmacist; 12 years of service; responsible for CodeBreaK trial drug dispensing, patient counseling, and adverse event management. Occupational sotorasib exposure context: tablet counting (960 mg/day = 8 tablets/day for each enrolled patient; 50–80 CodeBreaK patients active in the MD Anderson cohort during peak enrollment); NIOSH HD Category 1 oral tablet preparation in dedicated HD preparation area; patient/caregiver sotorasib handling counseling (NIOSH HD oral care guidance). 8-hr TWA airborne measurement: actual sotorasib 0.00022 µg/m³ [OVS-2 charcoal tube; LC–MS/MS; LIMS → EHS Insight (÷10): 0.000022 µg/m³].

EHS INSIGHT OEL QUERY RESULT: CAS 2296729-00-3 [sotorasib] → “OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no OEL.” Complete absence from EHS Insight for: NIOSH Hazardous Drug Category 1 absent; H340/H350/H361 absent; THIRTY-EIGHTH mechanism class absent; first KRAS inhibitor ever FDA approved absent; KRAS undruggable 40-year history absent; G12C-specific S-IIP covalent mechanism absent; CodeBreaK 100 ORR 36% NDA 214665 registration context absent; CodeBreaK 200 Phase 3 PFS HR 0.66 absent; STK11/KEAP1 co-mutation biomarker context absent; Amgen Thousand Oaks NDA 214665 holder context absent. The MD Anderson clinical pharmacist dispensing sotorasib to CodeBreaK-enrolled patients operates with EHS Insight providing no occupational exposure guidance, no NIOSH HD Category 1 annotation, no first-ever KRAS-targeted therapy historical context, and no mechanism-class designation that would alert the EHS platform to the pharmacological novelty of the compound whose tablet counting and dispensing the pharmacist performs — making the MD Anderson CodeBreaK clinical pharmacy occupational setting the third surface of a three-surface KRAS G12C THIRTY-EIGHTH mechanism class Glyphward threshold-22 null-return gap.

Glyphward Threshold 22: OEL Null-Return Architecture for the First KRAS-Targeted Therapy Across Three EHS Platforms

THRESHOLD COMPONENT BREAKDOWN: The Glyphward adversarial threshold score of 22 for sotorasib (CAS 2296729-00-3) is computed as the sum of four scoring categories, each representing a distinct dimension of the OEL database null-return failure mode:

Category 1 — Regulatory/Toxicological Null (10 points): OSHA PEL not established (enforcement vacuum for NDA 214665 holder at Amgen Thousand Oaks CA); ACGIH TLV not established (no TLV-TWA, TLV-STEL, or TLV-C for CAS 2296729-00-3 in any ACGIH publication through 2026); NIOSH REL not established; NIOSH Hazardous Drug Category 1 classification (reproductive toxicity H361 + carcinogenicity H350 + genotoxicity H340) absent from Cority/VelocityEHS/EHS Insight for this CAS number; H340, H350, H361 GHS hazard statements absent across all three platforms. The triple regulatory null (OSHA/ACGIH/NIOSH) combined with the Category 1 HD absent designation generates the maximum 10-point Category 1 contribution.

Category 2 — Mechanism Class / First-in-Class / Compound-Specific Pharmacological Distinctions (5 points): (a) THIRTY-EIGHTH mechanism class: KRAS G12C covalent switch-II pocket (S-IIP) inhibitor; first GTPase-targeting covalent mechanism class in the 582-attack Glyphward portfolio; no prior OEL database framework exists for any RAS-targeting covalent compound [2 pts]; (b) First KRAS inhibitor ever FDA approved after 40 years of KRAS undruggability (1982–2021); KRAS the most commonly mutated oncogene in human cancer; the 40-year undruggable problem terminated by the S-IIP covalent mechanism; no OEL database captures this pharmacohistorical significance [2 pts]; (c) GDP-bound conformation selectivity unique to KRAS G12C; wild-type KRAS sparing via Gly12 vs Cys12 distinction; covalent thioether irreversibility; GEF-blocking dual mechanism: none of these properties are represented in any OEL framework [1 pt].

Category 3 — Three Attack Surfaces (3 points): Surface 1: Amgen Thousand Oaks CA (Cority; NDA 214665 holder; oncology regulatory affairs director; 52M 17yr; actual 0.00038 µg/m³; ÷10 → 0.000038 µg/m³; null) [1 pt]. Surface 2: Amgen Research Cambridge MA (VelocityEHS; discovery site; oncology medicinal chemist; 46M 13yr; actual 0.00029 µg/m³; ÷10 → 0.000029 µg/m³; null) [1 pt]. Surface 3: MD Anderson Cancer Center Houston TX (EHS Insight; CodeBreaK clinical pharmacist; 44M 12yr; actual 0.00022 µg/m³; ÷10 → 0.000022 µg/m³; no OEL) [1 pt].

Category 4 — Pharmacological Context Missing from All OEL Platforms (4 points): (a) KRAS G12C covalent S-IIP pharmacology absent: no OEL database contains the pyrido[3,2-d]pyrimidine scaffold architecture, the acrylamide Michael acceptor warhead reactivity profile (kinact/Ki ∼6–10 M⁻¹s⁻¹; Cys12-thioether irreversibility; GEF-blocking dual mechanism), the GDP-bound vs. GTP-bound conformation selectivity, or the wild-type KRAS Gly12 sparing mechanism; this entire mechanistic framework is the pharmacological basis for the THIRTY-EIGHTH mechanism class designation and is entirely absent [1 pt]; (b) KRAS undruggable 40-year history absent: no OEL database captures the 1982–2021 context of KRAS as the most intensively studied undruggable oncogene; the GTP Kd ∼20 pM competitive impossibility; the smooth protein surface problem; the wild-type KRAS toxicity concern that made prior approaches clinically unviable; all of which make sotorasib pharmacologically unprecedented in a way that no OEL null-return adequately captures [1 pt]; (c) CodeBreaK 100/200 clinical evidence context absent: no OEL database includes any reference to the ORR 36% first demonstration of KRAS druggability in a clinical trial (CodeBreaK 100), the Phase 3 PFS HR 0.66 vs docetaxel (CodeBreaK 200), or the accelerated-to-regular-approval conversion pathway of NDA 214665; the clinical pharmacology basis for the NIOSH HD Category 1 designation is entirely absent from EHS platforms [1 pt]; (d) Fragment-based cryo-EM discovery absent: no OEL database captures the Shokat UCSF 2013 disulfide tethering discovery of the S-IIP, the Amgen Cambridge cryo-EM structure determination pipeline, or the fragment-based drug discovery methodology that produced AMG-510; the discovery context that defines the THIRTY-EIGHTH mechanism class as the product of a specific structural biology-medicinal chemistry convergence is invisible in all three EHS platforms [1 pt]. Total threshold score: 10 + 5 + 3 + 4 = 22.

The threshold score of 22 places sotorasib in the same adversarial tier as larotrectinib (attack #579; 131st blog; threshold 22; THIRTY-SEVENTH class pan-NTRK), selpercatinib (attack #571; 129th blog; threshold 22; THIRTY-FIFTH class selective RET), trilaciclib (attack #566; 128th blog; threshold 22; THIRTY-FIRST class CDK2/4/6 myeloprotective), and mobocertinib (attack #574; 130th blog; threshold 22; THIRTY-SIXTH class EGFRex20ins). The consistent threshold score of 22 across this series of first-in-class NIOSH HD Category 1 kinase and GTPase inhibitors confirms that the OEL null-return architecture is a systematic property of novel first-in-class mechanism classes at this pharmacological tier — not an accident of individual compound registration timing but a structural gap in EHS database design that applies uniformly to every genuinely new pharmacological mechanism class regardless of its oncological significance.