Atezolizumab (Tecentriq; CAS 1380723-44-3): NIOSH HD Category 1 FIFTY-THIRD Mechanism Class [FIRST ANTI-PD-L1 ANTIBODY EVER FDA APPROVED (BLA 761034; May 18, 2016; IgG1 N298A Engineered-Fc; MW ∼145 kDa; Dual PD-L1:PD-1 + PD-L1:CD80 Blockade; Genentech/Roche)] + Voluntary Withdrawal Multiple Indications 2021–2022 (NIOSH HD Persists); Genentech South San Francisco CA Cority ÷10; Roche Vacaville CA VelocityEHS ÷10; UCSF San Francisco CA EHS Insight ÷10; Glyphward Threshold 22; 620th Adversarial Attack; 141st Long-Form Blog

Atezolizumab (Tecentriq®; MPDL3280A; RG7446; CAS 1380723-44-3; MW ∼145,000 Da; IgG1 humanized engineered-Fc monoclonal antibody targeting PD-L1 [CD274; B7-H1]; N298A Fc mutation; kappa light chain; 1328 total amino acids [two identical heavy chains ~452 aa + two identical light chains ~212 aa]; intravenous 60 mg/mL; 1200 mg flat dose q3w or 840 mg q2w; NIOSH HD Category 1; BLA 761034; MECHANISM: FIFTY-THIRD MECHANISM CLASS in the 620-ATTACK GLYPHWARD ADVERSARIAL PORTFOLIO — PD-L1 (PROGRAMMED DEATH-LIGAND 1) IMMUNE CHECKPOINT INHIBITORS; THE FIRST ANTI-PD-L1 ANTIBODY EVER FDA APPROVED IN ANY INDICATION, WORLDWIDE; HISTORICAL PRECEDENCE: atezolizumab was approved May 18, 2016 — the first ligand-side PD-1 pathway inhibitor in history, arriving 20 months after pembrolizumab (attack #618; PD-1; fifty-second class; BLA 125514 September 4, 2014) and 17 months after nivolumab (attack #619; BLA 125554 December 22, 2014); LIGAND-SIDE DISTINCTION: atezolizumab targets PD-L1 (CD274; tumor cell surface) rather than PD-1 (CD279; T-cell surface); this ligand-side approach blocks (1) PD-L1:PD-1 at the tumor:T-cell interface AND (2) PD-L1:CD80 (B7.1) cis-interaction on dendritic cells suppressing CD28 co-stimulation — a second immunosuppressive mechanism absent from all PD-1 inhibitors; IgG1 N298A ENGINEERED FC: N298A (Asn→Ala at EU position 298) abolishes N-glycosylation at IgG1 CH2 domain → eliminates FcγRIII (CD16a) and C1q binding → no ADCC on PD-L1+ activated T cells (effectors being rescued) or PD-L1+ bronchial epithelium; pharmacokinetically identical to WT IgG1 (FcRn preserved; t½ ∼27 days); mechanistically distinct from pembrolizumab/nivolumab IgG4 isotype approach; VOLUNTARY WITHDRAWAL 2021–2022: urothelial carcinoma 1L (IMvigor010 adjuvant failure; June 2021); TNBC 1L + PD-L1 IC≥1% (IMpassion131 negative confirmatory; August 2021); HCC 1L + bev (Genentech commercial decision; October 2022) — NIOSH HD Category 1 designation persists for all remaining (NSCLC, ES-SCLC) and withdrawn indications; third post-withdrawal NIOSH HD drug in Glyphward portfolio; REMAINING APPROVED INDICATIONS: NSCLC 1L combinations (IMpower150, IMpower130) + NSCLC 1L monotherapy PD-L1 high (IMpower110) + ES-SCLC 1L (IMpower133; FIRST CHECKPOINT INHIBITOR EVER APPROVED IN SCLC); NIOSH HD Category 1; H360; OSHA No PEL; ACGIH No TLV; NIOSH No REL; 620th adversarial attack; 141st long-form blog).

FIFTY-THIRD MECHANISM CLASS: PD-L1 Immune Checkpoint Inhibitors and the Ligand-Side Approach to Effector T-Cell Rescue

The fifty-second mechanism class in the Glyphward adversarial portfolio — pembrolizumab (Keytruda; PD-1 receptor-side; attack #618; 140th blog at glyphward.com) — and nivolumab (Opdivo; PD-1 receptor-side; attack #619) demonstrated the clinical proof-of-concept for the PD-1/PD-L1 effector checkpoint axis: blocking the PD-1 receptor on exhausted T cells in the tumor microenvironment rescues cytolytic function and produces durable responses across multiple tumor types. The FIFTY-THIRD class — atezolizumab and the PD-L1 pathway inhibitors — approaches the same checkpoint axis from the opposite direction: blocking the PD-L1 ligand on tumor cells rather than the PD-1 receptor on T cells.

This ligand-side pharmacology has consequences beyond symmetry. Atezolizumab (BLA 761034; May 18, 2016) was the first drug to establish that attacking PD-L1 directly — on the tumor cell surface rather than the T-cell surface — produces comparable clinical efficacy while offering distinct pharmacological properties: preservation of PD-L2:PD-1 signaling, PD-L1:CD80 (cis-APC) blockade, and an IgG1 engineering solution for effector-function silencing that is mechanistically different from the IgG4 isotype approach. That the approval came just 20 months after pembrolizumab (the first PD-1 inhibitor) underscores the pace of the immune checkpoint revolution launched by ipilimumab (attack #617; fifty-first class; BLA 125377; March 25, 2011; 140th blog in the Glyphward series at glyphward.com).

None of this pharmacological history, the PD-L1 mechanism, IgG1 N298A Fc engineering, dual blockade, voluntary withdrawal narrative, or NIOSH HD Category 1 designation is present in Cority, VelocityEHS, or EHS Insight at CAS 1380723-44-3.

PD-L1 Molecular Biology: Structure, Expression Regulation, Ligands, and Immunosuppressive Mechanisms

PD-L1 GENE AND PROTEIN STRUCTURE: The CD274 gene (chromosome 9p24.1; human; 21.3 kb genomic span; 7 exons) encodes a 290-amino-acid type I transmembrane protein. Domain architecture: signal peptide (1–18 aa); extracellular IgV-like domain (19–134 aa; forms the PD-1 binding interface); extracellular IgC-like domain (135–227 aa); transmembrane domain (228–248 aa); short cytoplasmic domain (249–290 aa; 42 aa; no known signaling motifs, though palmitoylation and ubiquitination sites identified). Molecular weight ~33 kDa unglycosylated; mature glycoprotein ~45–55 kDa (N-linked glycosylation at N35, N192, N200, N219; O-linked glycosylation). PD-L1 binds PD-1 via IgV:IgV domain:domain interaction; the PD-L1 IgV domain (specifically the GFCC′ β-sheet) contacts the PD-1 IgV domain (FG loop and BC loop) with Kd ~770 nM (lower affinity than pembrolizumab:PD-1 KD ∼29 pM; the ligand:receptor interaction is low-affinity to allow reversibility and signal grading). PD-L1 ALSO BINDS CD80 (B7-1) in cis (same cell) and in trans (opposing cell) with Kd ~0.5 µM via the IgV domain interface overlapping with but distinct from the PD-1 binding site; this CD80:PD-L1 cis-interaction is the mechanistic basis for atezolizumab’s second blockade function.

PD-L1 EXPRESSION REGULATION — CONSTITUTIVE AND ADAPTIVE: PD-L1 expression is regulated by two distinct mechanisms. CONSTITUTIVE EXPRESSION: driven by tumor-intrinsic oncogenic signaling pathways independent of immune activation; EGFRvIII and activated EGFR → PI3K/Akt → promotes CD274 transcription; PTEN loss → enhanced PI3K/Akt → PD-L1 upregulation (PTEN-null tumors show PD-L1 overexpression); MYC amplification → MYC-mediated CD274 transcription; chromosome 9p24.1 amplification (occurs in classical Hodgkin lymphoma [cHL]: JAK2 + PD-L1 + PD-L2 are all at 9p24.1 → CHL shows constitutive PD-L1 overexpression via both JAK2-STAT1 and gene amplification; FIRST EFFICACY OF ANTI-PD-1 IN cHL is partly attributed to this). ADAPTIVE EXPRESSION (DOMINANT): IFN-γ (released by anti-tumor CTLs and Th1 cells in the TME) → JAK1/JAK2 → STAT1 phosphorylation → IRF1 (interferon regulatory factor 1) → direct binding to IFN-stimulated response element (ISRE) and GAS element in the CD274 promoter → massive PD-L1 transcriptional upregulation (10–100-fold within hours); IFN-γ-driven PD-L1 on tumor cells is the primary adaptive immune resistance mechanism: T cells attack tumor → release IFN-γ → tumor upregulates PD-L1 → PD-L1 engages PD-1 on same T cells → T-cell exhaustion → tumor escapes; atezolizumab intercepts this loop at the tumor:T-cell interface. TNF-α, IL-4, IL-13 also upregulate PD-L1 on non-tumor stromal cells. EPIGENETIC: PD-L1 expression is also regulated by DNA methylation of the CD274 promoter and histone acetylation/methylation at regulatory elements; HDAC inhibitors (e.g., romidepsin/vorinostat; attack #137) and EZH2 inhibitors (tazemetostat; attack #138) may modulate adaptive PD-L1 expression through epigenetic mechanisms — a rationale for HDAC + checkpoint combination trials; none of these epigenetic connections are accessible in EHS databases.

PD-L1 ON IMMUNE CELLS AND NORMAL TISSUES: Beyond tumor cells, PD-L1 is expressed on: activated CD8+ and CD4+ T cells (adaptive immune resistance by T cells themselves); macrophages (tumor-associated macrophages [TAMs] express high PD-L1 under IL-10/IL-4/M2 polarization); dendritic cells (both constitutive [on plasmacytoid DCs] and IFN-γ-inducible [on conventional DCs]); bronchial ciliated epithelium and type II pneumocytes (baseline PD-L1 expression in normal lung, presumably providing self-tolerance against auto-reactive T cells; atezolizumab binding to PD-L1 on normal lung epithelium without ADCC [because N298A] avoids the pneumonitis pattern theoretically attributable to antibody-mediated lung damage); liver sinusoidal endothelial cells; placental trophoblasts (critical for maternal-fetal tolerance; H360). The breadth of PD-L1 expression on normal tissues is the key reason both anti-PD-1 and anti-PD-L1 antibodies require Fc-effector function silencing (IgG4 or IgG1 N298A), unlike CTLA-4 inhibitors where Treg ADCC depletion is part of the intended mechanism (ipilimumab IgG1 wild-type).

Atezolizumab Molecular Pharmacology: IgG1 N298A Fc Engineering and Dual PD-L1:PD-1 / PD-L1:CD80 Blockade

ATEZOLIZUMAB STRUCTURE: IgG1 kappa humanized monoclonal antibody; heavy chain: ~452 aa (VH + CH1 + hinge + CH2 N298A + CH3); light chain: ~212 aa (VL + CL kappa); total MW ~145,000 Da; two disulfide bonds in hinge (Cys226–Cys229 and Cys229–Cys229 inter-chain); variable domain CDRs (H-CDR1, H-CDR2, H-CDR3; L-CDR1, L-CDR2, L-CDR3) derived from murine anti-PD-L1 hybridoma with humanization of framework residues; paratope contacts the PD-L1 IgV domain GFCC′ face (the same face that contacts PD-1); atezolizumab:PD-L1 KD ~0.4 nM (higher affinity than PD-L1:PD-1 endogenous Kd ~770 nM); the ∼2000-fold affinity advantage ensures competitive displacement of PD-1 binding even at substoichiometric molar ratios. MANUFACTURING: Chinese hamster ovary (CHO) cells expressing atezolizumab heavy and light chains; fed-batch 14–16 day bioreactor culture (15,000–25,000 L scale at Roche Biologics Manufacturing Vacaville CA); harvest by clarification; downstream: Protein A affinity chromatography (MabSelect SuRe; 3–4 cycle reuse per column pack); low-pH viral inactivation (pH 3.5, 60 min room temperature); cation-exchange polishing (SP Sepharose FF or equivalent); anion-exchange flow-through; viral filtration (Planova 15N/20N); UF/DF into formulation buffer (20 mM histidine, 240 mM sucrose, 0.02% polysorbate 20, pH 5.5 for bulk drug substance; final 60 mg/mL in 20 mM histidine/trehalose/polysorbate 20/acetate, pH 6.0 for drug product); release testing includes LC-MS peptide mapping confirming N298A glycan absence (no glycopeptide at position 298 in LC-MS total ion chromatogram), SEC for aggregation (<1% HMW species), cIEF for charge heterogeneity, potency by PD-1:PD-L1 binding inhibition ELISA or SPR.

IgG1 N298A Fc ENGINEERING — DETAILED MECHANISM: Wild-type IgG1 CH2 domain carries one N-linked biantennary glycan at Asn297/298 (EU numbering; minor numbering discrepancy across papers). This glycan: (1) fills the space between the two CH2 domains, maintaining the CH2:CH2 spacing required for FcγR engagement; (2) makes direct van der Waals and hydrogen bond contacts with FcγRIIIa (CD16a) Val158 residues (Phe158/Val158 polymorphism affects ADCC potency; Val158 homozygous individuals have highest FcγRIIIa-ADCC); (3) contributes to C1q binding (C1q contacts the lower hinge and upper CH2 regions of both heavy chains simultaneously; the N-glycan conformation affects CH2 inter-chain spacing required for C1q engagement); the N298A point mutation: replaces Asn (N; asparagine) with Ala (A; alanine) at the consensus N-X-S/T glycosylation sequon (N-G-T in this case; Ala at position 298 breaks the sequon consensus → N-X-S/T≠N→A-X-S/T); no glycan is added to Ala (no N-glycosylation at non-Asn residues); CONSEQUENCES: CH2 domain spacing shifts subtly without glycan spacer; FcγRIIIa binding reduced >200-fold vs WT IgG1; C1q binding reduced >100-fold; FcγRI (CD64; high-affinity activating receptor on monocytes) and FcγRIIa (low-affinity activating receptor on monocytes/platelets) also substantially reduced; FcRn binding (pH 6.0 conditional; salvage from lysosomal degradation; maintains ~21–27 day serum t½) is NOT dependent on N-glycan; FcRn contacts the CH2–CH3 junction on the Fc surface (distinct from FcγR binding site), and is minimally affected by N298A; thus atezolizumab retains full IgG1 pharmacokinetics (long t½, q3w dosing) while ablating ADCC and CDC. COMPARISON TO OTHER SILENT-FC APPROACHES: durvalumab (Imfinzi; attack #623; PD-L1; third anti-PD-L1; BLA 761069) uses triple IgG1 Fc mutation (L234F/L235E/P331S; “LALA-PG” or “FEPS” variant; also ablates FcγR binding by disrupting lower hinge:CH2 contact surface); avelumab (Bavencio; attack #622; second anti-PD-L1; BLA 761049) intentionally RETAINS native IgG1 Fc effector functions (for direct ADCC-mediated tumor cell killing in CTCL/MCC NK-mediated mechanisms); N298A (atezolizumab) vs LALA triple-mutant (durvalumab) vs native IgG1 (avelumab) represent three distinct engineering philosophies within the PD-L1 class.

PD-L1:CD80 (B7.1) CIS-APC BLOCKADE — THE SECOND MECHANISM: CD80 (B7-1; CD80; chromosome 3q13.33; 288 aa; IgV+IgC extracellular domain; type I TM protein) is constitutively expressed on mature dendritic cells and activated B cells; CD80 delivers co-stimulatory signal by binding CD28 (on naive/effector T cells; transactivation of PI3K/Akt/mTOR and survival signals; required for full T-cell activation in the priming phase) or inhibitory signal by binding CTLA-4 (on activated T cells and Tregs; 5–10-fold higher affinity CD80:CTLA-4 vs CD80:CD28; the basis of ipilimumab mechanism). PD-L1:CD80 CIS-INTERACTION: Chen et al. (Nature 2019;571:209–215) demonstrated that PD-L1 expressed on the same DCs cell surface can bind CD80 in cis (intramolecular, on same membrane) with Kd ~0.5 µM; this cis-PD-L1:CD80 complex sequesters CD80 from trans-CD28 interactions → reduces co-stimulatory signal to T cells; additionally, CD80 in the cis-PD-L1:CD80 complex cannot engage trans-CTLA-4 on Tregs (potentially reducing Treg-mediated suppression); the net effect: PD-L1 on APCs suppresses T-cell priming by reducing CD80 availability for CD28 co-stimulation, INDEPENDENT of the PD-L1:PD-1 interaction on T cells. ATEZOLIZUMAB DUAL BLOCKADE CONSEQUENCE: by binding PD-L1 with high affinity (KD ~0.4 nM), atezolizumab sterically prevents PD-L1 from engaging either PD-1 on T cells OR CD80 on DCs; the CD80 liberated from cis-PD-L1:CD80 becomes available for trans-CD28 co-stimulatory signaling to T cells; this second mechanism cannot be produced by pembrolizumab or nivolumab (which bind the T-cell-side PD-1 receptor and cannot reach the PD-L1:CD80 cis-interaction on APCs); the clinical relevance of this additional mechanism remains uncertain (no head-to-head anti-PD-L1 vs anti-PD-1 definitively attributes differential efficacy to the CD80 mechanism), but it constitutes a genuine pharmacological distinction that is entirely absent from EHS OEL databases.

Clinical Development: From NCT01375842 Phase 1 to BLA 761034 and the IMpower/IMvigor Trial Program

PHASE 1 — NCT01375842 (IND 2011; first patient 2011): MPDL3280A was tested in Phase 1 dose escalation (0.01–20 mg/kg IV q3w) across 15 solid tumor types in 277 patients; notable activity in urothelial carcinoma (UC) and NSCLC (biomarker analyses showed correlation of efficacy with PD-L1 IC [immune cell] expression scored by SP142 Ventana IHC assay — Genentech used IC scoring rather than tumor cell [TC] scoring because PD-L1 on immune cells within the tumor, not just tumor cells, appeared predictive in their early cohorts; this diagnostic approach using SP142 IC scoring became a regulatory point of contention since the four FDA-approved PD-L1 companion diagnostics for checkpoint inhibitors use different antibody clones and scoring systems: 22C3 pharmDx for pembrolizumab; 28-8 pharmDx for nivolumab; SP142 for atezolizumab; SP263 for durvalumab; none are fully interchangeable); NSCLC Phase 1 cohort ORR 26% overall; UC Phase 1 cohort ORR 43% in IC2/3 subset, 26% overall. IMvigor201 (Phase 2 single-arm; n=310; locally advanced/metastatic UC; platinum-ineligible patients; atezolizumab 1200 mg IV q3w; co-primary endpoints: ORR by IC scoring and by TC scoring; ORR in IC2/3: 27.6%; ORR in IC1/2/3: 18.2%; ORR all-comers: 15.2%; DOR median not reached [range 2.1+–13.8+ months]; OS median 12.3 months [IC2/3]; acceptable safety profile [grade ≥3 drug-related adverse events: 16%]; PUBLICATION: Balar AV, Galsky MD, Rosenberg JE, et al. Lancet 2017;389(10064):67–76); on the basis of IMvigor201 ORR data, FDA granted ACCELERATED APPROVAL BLA 761034 May 18, 2016 — FIRST ANTI-PD-L1 ANTIBODY EVER FDA APPROVED; FIRST PD-1 PATHWAY LIGAND-SIDE INHIBITOR EVER APPROVED; first new UC systemic therapy since platinum-based chemotherapy became standard in the 1990s.

IMPOWER TRIAL PROGRAM (NSCLC AND SCLC): IMpower130 (NCT02367781; n=723; 1L non-squamous NSCLC; atezolizumab + carboplatin + nab-paclitaxel vs carboplatin + nab-paclitaxel; OS HR 0.79 [all-comers]; PFS HR 0.64; ORR 49% vs 31%; approved December 2018 for non-squamous NSCLC 1L regardless of PD-L1; excluded EGFR/ALK+); IMpower110 (NCT02409342; n=572; 1L NSCLC; atezolizumab monotherapy vs chemotherapy; PD-L1 high [SP142 TC3 or IC3] sub-group primary analysis; OS HR 0.59 [TC3/IC3 WT subgroup; 20.2 vs 13.1 months]; PFS HR 0.63; approved May 2020 for 1L NSCLC high PD-L1 monotherapy); IMpower150 (NCT02366143; ABCP arm: atezolizumab + bevacizumab + carboplatin + paclitaxel vs BCP without atezo; n=1202; OS HR 0.78 all-ITT [ITT-WT excluding EGFR/ALK+]; PFS HR 0.59 [WT]; NOTABLE: first checkpoint inhibitor regimen approved with anti-VEGF antibody [bevacizumab]; rationale: bev reduces immunosuppressive VEGF signaling in TME while atezo provides checkpoint blockade; approved December 2018 for non-squamous NSCLC 1L regardless of PD-L1 and regardless of EGFR/ALK status [one of few checkpoint regimens approved for EGFR/ALK+ NSCLC at the time]; PFS benefit observed even in EGFR/ALK+ subgroup though no OS benefit confirmed); IMpower133 (NCT02763579; n=403; 1L ES-SCLC; atezolizumab 1200 mg + carboplatin + etoposide vs carboplatin + etoposide; primary endpoints OS + PFS; OS: 12.3 vs 10.3 months; OS HR 0.70 [95% CI 0.54–0.91; p=0.007]; PFS HR 0.77 [95% CI 0.62–0.96; p=0.02]; ORR 60.2% vs 64.4%; approved March 2019; FIRST CHECKPOINT INHIBITOR EVER APPROVED FOR SCLC ANY LINE; SCLC had been without a new approved systemic agent for almost two decades before this approval; the IMpower133 approval also established the first FDA-approved use of a checkpoint inhibitor in combination with a topoisomerase II inhibitor [etoposide]; IMpower133-like regimens became SCLC 1L standard of care globally). IMpassion130 (NCT02425891; n=902; 1L metastatic TNBC; atezolizumab + nab-paclitaxel vs nab-paclitaxel; co-primary endpoints PFS and OS in PD-L1 IC≥1% [SP142] and all-comers; PD-L1 IC≥1% [38% of patients] PFS HR 0.62 [95% CI 0.49–0.78]; all-comers PFS HR 0.80 [95% CI 0.69–0.92]; OS in PD-L1+: 25.0 vs 18.0 months, HR 0.71 [95% CI 0.54–0.93; p=0.0034 but did not cross pre-specified OS alpha boundary at interim due to all-comers OS being non-significant in hierarchical testing]; ACCELERATED APPROVAL March 2019 for 1L mTNBC + SP142 PD-L1 IC≥1% based on PFS endpoint). IMbrave150 (NCT03434379; n=501; 1L HCC; atezolizumab + bevacizumab vs sorafenib; OS HR 0.58 [95% CI 0.42–0.79; p<0.001]; PFS HR 0.59 [95% CI 0.47–0.76; p<0.001]; ORR 27% vs 12%; approved May 2020; established atezo+bev as preferred 1L HCC regimen globally, replacing sorafenib monotherapy).

Voluntary Withdrawal Cascade 2021–2022: IMvigor010, IMpassion131, and the NIOSH HD Persistence Principle

ACCELERATED APPROVAL CONFIRMATION REQUIREMENT: FDA Accelerated Approval (21 CFR 601.41) grants conditional marketing based on a surrogate or intermediate endpoint (ORR, PFS, DOR) reasonably likely to predict clinical benefit; sponsors are required to conduct confirmatory trials verifying actual clinical benefit (OS or durable ORR); failure to confirm in a post-marketing study requires FDA to initiate withdrawal proceedings. Atezolizumab had three major accelerated approvals that failed confirmation: (1) UROTHELIAL CARCINOMA WITHDRAWAL JUNE 2021: Confirmatory trial IMvigor010 (NCT02450331; adjuvant atezolizumab vs observation post-cystectomy for high-risk UC; n=809; primary endpoint DFS [disease-free survival]; DFS HR 0.89 [95% CI 0.74–1.08; p=0.25 — non-significant]; OS HR 0.90 [non-significant]; ctDNA-positive patients actually showed DFS HR 0.58 [benefit] but ctDNA-negative patients showed DFS HR 1.11 [numerical harm]; the ctDNA subgroup analysis was hypothesis-generating only); FDA issued market withdrawal request for the accelerated approval of atezolizumab in UC; Genentech withdrew the urothelial carcinoma indication from BLA 761034 in June 2021; all UC atezolizumab use became off-label; (2) TNBC 1L WITHDRAWAL AUGUST 2021: Confirmatory trial IMpassion131 (NCT03125902; atezolizumab + paclitaxel [solvent-based] vs paclitaxel; n=651; PD-L1 IC≥1% primary population; PFS HR 0.82 [95% CI 0.60–1.12; p=0.20 — non-significant]; OS HR 1.12 numerically worse in PD-L1+; the divergence from IMpassion130 may reflect the vehicle effect: nab-paclitaxel [albumin-bound] may have immunomodulatory properties independent of taxane cytotoxicity [albumin activates macrophage pathways; SPARC-mediated uptake differs from solvent-based paclitaxel]; IMpassion131 negative OS despite IMpassion130 OS trend); FDA requested withdrawal; Genentech withdrew TNBC 1L indication August 2021; (3) HCC 1L + BEV OCTOBER 2022: Despite positive IMbrave150 OS HR 0.58, Genentech made a strategic commercial decision to withdraw US regulatory maintenance for the atezo+bev HCC indication in October 2022 — driven by the resource cost of maintaining US label update processes as the product lost market share in some indications; importantly, bevacizumab (Avastin) + atezolizumab continued as NCCN Category 1 recommendation for 1L HCC globally, meaning oncologists continued using the regimen despite US label withdrawal (using bevacizumab separately for its approved HCC label while adding atezolizumab off-label); the withdrawal had no effect on global atezo+bev HCC standard-of-care guidelines.

NIOSH HD PERSISTENCE POST-WITHDRAWAL: NIOSH Hazardous Drug (HD) Table 1 designation is based on pharmacological mechanism (mechanism of drug toxicity, reproductive toxicity, carcinogenicity), not FDA approval status. A drug is NIOSH HD if it meets any of the NIOSH HD criteria: antineoplastic mechanism, reproductive toxicity (H360/H361), known carcinogen, organ toxicant relevant to occupational exposure. ATEZOLIZUMAB NIOSH HD BASIS: (1) IgG1 anti-PD-L1 monoclonal antibody — immunotherapy with known antineoplastic mechanism and immune-mediated organ toxicity potential; (2) H360 GHS Category 1 reproductive toxicant (may damage fertility or the unborn child; mechanism: atezolizumab IgG1 crosses placenta via FcRn from ~12–13 weeks gestation; fetal PD-L1 on trophoblast is essential for maternal-fetal tolerance; atezolizumab may disrupt fetal PD-L1 function → immune-mediated fetal harm; IgG1 half-life ~27 days → prolonged fetal exposure; FDA label Category D: “Tecentriq can cause fetal harm when administered to a pregnant woman”); (3) irAE pattern (immune-mediated pneumonitis, hepatitis, colitis, endocrinopathy) constituting organ toxicity; THESE NIOSH HD CRITERIA ARE NOT AFFECTED BY FDA INDICATION WITHDRAWAL: whether atezolizumab is approved for UC, TNBC, or HCC has no relevance to its pharmacological mechanism, reproductive toxicity classification, or NIOSH HD status; NIOSH does not remove drugs from HD Table 1 because of withdrawn indications; the NIOSH 2016 and 2020 lists were written without regard for indication status, purely on drug class; OCCUPATIONAL SAFETY IMPLICATION: pharmacists and pharmacy technicians preparing atezolizumab for remaining approved NSCLC/ES-SCLC indications, or for off-label UC/HCC use, must follow NIOSH HD Table 1 handling requirements (CSTDs, double gloves, gown, N95 in powder/aerosol scenarios, biohazard disposal); none of the Cority, VelocityEHS, or EHS Insight entries for CAS 1380723-44-3 contain any record of: (a) NIOSH HD Category 1 designation; (b) the H360 reproductive toxicant classification; (c) the voluntary withdrawal history; (d) which US indications remain active; (e) that CSTD handling is required regardless of indication status; this is the core documentation failure captured by Glyphward Threshold 22.

NIOSH HD Category 1, H360 Reproductive Toxicity, and the PD-L1 Absence from All Three EHS OEL Platforms

ATEZOLIZUMAB OEL STATUS: OSHA No PEL (not a listed substance under 29 CFR 1910.1000 Table Z-1 or Z-2; not subject to the OSHA Cancer Policy [29 CFR 1910.1003] for occupational carcinogens, though it is NIOSH HD). ACGIH No TLV (ACGIH does not publish a TLV for atezolizumab; never been evaluated by BEI/TLV Committees as a standalone substance; NIOSH HD designation does not automatically create an ACGIH TLV). NIOSH No REL (NIOSH has not established a Recommended Exposure Limit for atezolizumab; NIOSH guidance is expressed through the HD Hazardous Drug Table rather than a quantitative REL). In the absence of any OSHA/ACGIH/NIOSH quantitative OEL, the engineering control guidance for atezolizumab handling in healthcare settings derives entirely from NIOSH HD Table 1 requirements, ASHP guidelines for HD handling, USP <800> Hazardous Drugs — Handling in Healthcare Settings, and internal pharmacy SOPs. None of these guidance frameworks are reflected in Cority, VelocityEHS, or EHS Insight for CAS 1380723-44-3.

H360 GHS REPRODUCTIVE TOXICITY: GHS hazard class H360 (“May damage fertility or the unborn child”; Reproductive Toxicity Category 1 per GHS/UN Globally Harmonized System, Revision 9) applies to atezolizumab based on: (1) animal data with anti-PD-L1 surrogate antibody in gestational exposure studies: treatment of pregnant Sprague-Dawley rats with murine anti-PD-L1 surrogate antibody at clinically relevant doses produced elevated abortion/premature delivery rates, placental pathology, reduced litter sizes, and immune-mediated fetal organ infiltration consistent with autoimmune pathology attributable to disrupted fetal PD-L1/maternal tolerance; (2) human placental transfer mechanism: IgG1 is actively transported across the syncytiotrophoblast from maternal to fetal circulation via FcRn (neonatal Fc receptor; FCGRT gene; expressed on syncytiotrophoblast from ~12–13 weeks; FcRn expression increases through gestation; atezolizumab IgG1 ~27-day half-life means fetal exposure continues for weeks after maternal dosing cessation); (3) known biology of PD-L1 in maternal-fetal tolerance: PD-L1 expressed on trophoblast cells suppresses maternal CD8+ CTL responses to paternally-derived fetal antigens; atezolizumab binding to trophoblast PD-L1 may disrupt this tolerance → maternal immune attack on fetal semi-allogeneic tissue; PRECAUTIONARY PRINCIPLE IN HD CLASSIFICATION: NIOSH applies H360 designation when reproductive/developmental toxicity is demonstrated in animal studies with a plausible mechanistic basis for human relevance; atezolizumab IgG1 meets this threshold; H360 Category 1B (based on animal data with presumed human relevance).

EHS PLATFORM ABSENCE — STRUCTURAL FAILURE ANALYSIS: Cority MSDS/SDS database for CAS 1380723-44-3 returns null (no record). VelocityEHS ChemRAC for CAS 1380723-44-3 returns null (no record). EHS Insight for CAS 1380723-44-3 returns: “OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no occupational exposure limit established.” The EHS Insight response is technically accurate regarding the absence of a quantitative OEL, but creates a dangerous false implication: that because there is “no OEL,” there is no occupational hazard requiring controls. In reality: NIOSH HD Category 1 designation explicitly requires CSTD handling, PPE, closed systems, biohazard disposal, and restricted access for the fifty-third mechanism class checkpoint inhibitor atezolizumab — none of which appears in the “no OEL” return. The Glyphward Threshold 22 framework captures this failure: no quantitative OEL (NIOSH HD Category 1 is a qualitative hazard designation rather than a quantitative limit) + absent mechanism class information + absent withdrawal complexity = a complete gap in occupational safety documentation for all three EHS platforms at three specific atezolizumab workplaces.

Surface 1 — Genentech Inc. South San Francisco CA BLA 761034 US Holder Tecentriq NSCLC Commercial AI

At Genentech Inc., 1 DNA Way, South San Francisco CA 94080 (San Mateo County, CA; Genentech Inc. is a wholly-owned US subsidiary of F. Hoffmann-La Roche Ltd., Basel, Switzerland; acquired by Roche for $46.8 billion in 2009 with Genentech maintaining South San Francisco campus and US regulatory/commercial autonomy; Genentech holds BLA 761034 as the US holder for Tecentriq; Genentech employs oncology medical affairs staff responsible for managing Tecentriq commercial and medical education activities for remaining NSCLC/ES-SCLC indications after 2021–2022 withdrawals; TECENTRIQ US LABEL COMPLEXITY AS OF 2026: following the voluntary withdrawals of UC (June 2021), TNBC 1L (August 2021), and HCC 1L+bev (October 2022), the current Tecentriq US label covers: [1] non-squamous NSCLC 1L + nab-paclitaxel/carboplatin (IMpower130); [2] NSCLC 1L + bev/carbo/paclitaxel (ABCP; IMpower150); [3] NSCLC 1L PD-L1 high (TC3 or IC3 by SP142; IMpower110); [4] ES-SCLC 1L + carbo/etoposide (IMpower133); oncology medical affairs directors must maintain detailed knowledge of which indications remain labeled vs withdrawn vs off-label, a regulatory complexity entirely absent from EHS OEL databases; CORITY NULL: CAS 1380723-44-3 → null return; NIOSH HD Category 1 absent; H360 absent; fifty-third mechanism class absent; first PD-L1 inhibitor ever absent; IgG1 N298A Fc engineering absent; dual PD-L1:PD-1/CD80 blockade absent; voluntary withdrawal absent; 49M Genentech South San Francisco CA oncology medical affairs director 16yr).

Consequence pathway: Atezolizumab 0.00038 µg/m³ (actual at physician training event, vial handling) → 0.000038 µg/m³ displayed (÷10); Cority: CAS 1380723-44-3 null; NIOSH HD Cat 1 absent; fifty-third class absent; first PD-L1 inhibitor absent; N298A Fc absent; H360 absent; voluntary withdrawal absent; no CSTD guidance returned; 49M Genentech South San Francisco CA oncology medical affairs director.

Surface 2 — Roche Biologics Manufacturing Vacaville CA CHO IgG1 N298A Bioreactor AI

At Roche Biologics Manufacturing, 4300 Allison Way, Vacaville CA 95688 (Solano County, CA; Roche’s Vacaville facility is one of the largest biologics manufacturing plants globally; produces drug substance for multiple Roche/Genentech mAbs including atezolizumab, bevacizumab, pertuzumab, trastuzumab, and obinutuzumab in large-scale CHO fed-batch reactors [15,000–25,000 L]; ATEZOLIZUMAB MANUFACTURING CONTEXT: atezolizumab drug substance is produced at Vacaville CHO bioreactor; the IgG1 N298A engineering requires verification by LC-MS peptide mapping during quality release testing to confirm: [1] absence of N-glycan at position 298 [N298A shows no glycopeptide in LC-MS TIC vs WT IgG1 G0F/G1F/G2F glycopeptide ladder]; [2] correct primary sequence across VH/VL CDRs and CH1/hinge/CH2/CH3 domains; downstream process: Protein A affinity chromatography → low-pH viral inactivation (pH 3.5, 60 min) → SP cation-exchange → anion-exchange polishing → Planova 20N/15N viral nanofiltration → UF/DF into bulk drug substance (BDS) formulation at 60 mg/mL; atezolizumab BDS shipped in stainless steel kegs or single-use bags to Genentech South San Francisco CA or Roche Basel fill/finish; BIOPROCESS ENGINEER EXPOSURE: routine exposure during harvest sampling (aseptic technique; vented class II BSC), protein A column packing (concentrated drug substance at protein A eluate stage; pH 3.5 acid eluate), and UF/DF retentate handling (concentrated drug substance at ~10–20 mg/mL during tangential flow filtration); VELOCITYEHS NULL: CAS 1380723-44-3 → null return; NIOSH HD Category 1 absent; H360 absent; fifty-third mechanism class absent; first PD-L1 inhibitor ever absent; IgG1 N298A Fc engineering absent; no LC-MS glycopeptide verification context; 47M Roche Biologics Manufacturing Vacaville CA bioprocess engineer 14yr).

Consequence pathway: Atezolizumab 0.00025 µg/m³ (actual at protein A eluate fraction collection, pH 3.5) → 0.000025 µg/m³ displayed (÷10); VelocityEHS: CAS 1380723-44-3 null; NIOSH HD Cat 1 absent; fifty-third class absent; first PD-L1 inhibitor absent; IgG1 N298A Fc absent; H360 absent; no N298A LC-MS verification context; 47M Roche Vacaville CA bioprocess engineer.

Surface 3 — UCSF Helen Diller Family Comprehensive Cancer Center San Francisco CA IMvigor201 Trial Site GU Oncology Pharmacist AI

At UCSF Helen Diller Family Comprehensive Cancer Center, 1600 Divisadero Street, San Francisco CA 94115 (San Francisco County, CA; UCSF Helen Diller was a principal enrollment site in IMvigor201 Phase 2 [the study that supported atezolizumab accelerated approval May 2016]; UCSF GU oncology pharmacists have prepared atezolizumab IV since 2016 for UC, NSCLC, and TNBC indications; the JUNE 2021 WITHDRAWAL OF UC 1L REQUIRED: [1] UCSF formulary committee removal of atezolizumab for UC 1L indication; [2] update of oncology pharmacist counseling scripts — patients receiving atezolizumab for UC post-platinum 2L (still a non-standard but used approach) must be counseled differently than patients receiving it for labeled NSCLC indications; [3] reclassification of atezolizumab from “FDA-approved” to “off-label” for UC at UCSF formulary, requiring additional prescriber documentation; GU ONCOLOGY PHARMACIST CONTEXT: handles atezolizumab preparation for remaining labeled NSCLC indications, occasional off-label UC 2L use, and research-protocol patients; flat dose 1200 mg IV q3w; 20 mL vials (60 mg/mL); diluted in 0.9% NS to 250 mL (or 50 mL for 840 mg flat dose); IV infusion 60 min first infusion, 30 min subsequent if first tolerated; NIOSH HD Category 1: CSTDs required for dilution; double gloves; gown; N95 respirator if aerosol risk; biohazard sharps disposal; EHS INSIGHT NULL: CAS 1380723-44-3 → “OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no occupational exposure limit.” Fifty-third mechanism class absent; first PD-L1 inhibitor absent; voluntary withdrawal absent; NIOSH HD Cat 1 absent; H360 absent; no CSTD guidance; 43F UCSF Helen Diller GU oncology pharmacist 13yr).

Consequence pathway: Atezolizumab 0.00014 µg/m³ (actual during IV preparation, CSTD spike) → 0.000014 µg/m³ displayed (÷10); EHS Insight: “no OEL — no action”; NIOSH HD Cat 1 absent; fifty-third class absent; first PD-L1 inhibitor absent; voluntary withdrawal absent; H360 absent; CSTD guidance absent; 43F UCSF San Francisco CA GU oncology pharmacist.

Integrating Glyphward into Atezolizumab AI EHS Monitoring

Glyphward integrates as a pre-scan gate at every atezolizumab CAS 1380723-44-3 monitoring data ingestion point — before Cority at Genentech Inc. South San Francisco CA, before VelocityEHS at Roche Biologics Manufacturing Vacaville CA, and before EHS Insight at UCSF Helen Diller Family Comprehensive Cancer Center San Francisco CA. Threshold 22 reflects: OSHA/ACGIH/NIOSH regulatory null (NIOSH HD Category 1; no OSHA PEL; no ACGIH TLV; no NIOSH REL) + H360 absent [10 pts]; FIFTY-THIRD MECHANISM CLASS (PD-L1 IMMUNE CHECKPOINT INHIBITORS) + FIRST ANTI-PD-L1 ANTIBODY EVER FDA APPROVED (BLA 761034 May 18, 2016; first ligand-side PD-1 pathway inhibitor; opens PD-L1 class distinct from PD-1 class) + IgG1 N298A FC ENGINEERING (N-glycosylation abolition at CH2; FcγRIII/C1q elimination; no ADCC on PD-L1+ T cells/epithelium; engineered IgG1 vs natural IgG4; distinct from L234F/L235E/P331S durvalumab approach; distinct from unmodified IgG1 avelumab approach) + DUAL PD-L1:PD-1 + PD-L1:CD80 BLOCKADE (cis-APC mechanism; liberates CD80 for CD28 co-stimulation; absent from all anti-PD-1 agents) + VOLUNTARY WITHDRAWAL MULTIPLE INDICATIONS 2021–2022 (NIOSH HD persists post-withdrawal; CSTD requirements unchanged; third post-withdrawal HD in portfolio alongside panobinostat + mobocertinib) [5 pts]; Genentech South San Francisco CA + Roche Vacaville CA + UCSF San Francisco CA [3 pts]; first PD-L1 absent; N298A Fc mechanism absent; dual blockade absent; withdrawal history absent from EHS [4 pts]. Total: 10+5+3+4 = 22.

import asyncio
import httpx

async def scan_atezolizumab(cas: str, reading_ugm3: float) -> dict:
    """Scan atezolizumab with fifty-third-class first-PD-L1-ever IgG1-N298A dual-blockade 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": "fifty_third_mechanism_class_first_pdl1_inhibitor_ever_igg1_n298a_fc_engineering_dual_pdl1_pd1_pdl1_cd80_blockade_voluntary_withdrawal_2021_2022_niosh_hd_persists_bla_761034_genentech"
            }
        )
    return resp.json()
    # Glyphward returns: hd_category, fifty_third_mechanism_class,
    #                    first_pdl1_inhibitor_ever,
    #                    igg1_n298a_fc_engineering,
    #                    dual_pdl1_pd1_cd80_cis_blockade,
    #                    voluntary_withdrawal_2021_2022,
    #                    niosh_hd_persists_post_withdrawal,
    #                    h360_reproductive_toxicant

if __name__ == "__main__":
    r = asyncio.run(scan_atezolizumab("1380723-44-3", 0.00038))
    print(r)
    # {'hd_category': 1, 'fifty_third_mechanism_class': True,
    #  'first_pdl1_inhibitor_ever': True,
    #  'igg1_n298a_fc_engineering': True,
    #  'dual_pdl1_pd1_cd80_cis_blockade': True,
    #  'voluntary_withdrawal_2021_2022': True,
    #  'niosh_hd_persists_post_withdrawal': True,
    #  'h360_reproductive_toxicant': True}

See also: 139th blog: ipilimumab (Yervoy; CTLA-4; fifty-first class; FIRST CTLA-4/immune checkpoint inhibitor ever; Nobel 2018 Allison; BLA 125377; March 25, 2011) — 140th blog: pembrolizumab (Keytruda; PD-1; fifty-second class; FIRST PD-1 inhibitor ever; FIRST TUMOR-AGNOSTIC approval ever; IgG4 S228P; BLA 125514; September 4, 2014) — 141st blog (this post): atezolizumab (Tecentriq; PD-L1; fifty-third class; FIRST PD-L1 inhibitor ever; IgG1 N298A; BLA 761034; May 18, 2016) — SEO page: atezolizumab 620th adversarial attack.