What is the thienopyrimidine scaffold mechanism of fedratinib (C27H36N6O3S; MW ~560.68 g/mol; CAS 936091-14-4), how does JAK2 selectivity (~35-fold: JAK2 IC50 ~3 nM; JAK1 IC50 ~105 nM) mechanistically differ from ruxolitinib's non-selective JAK1/2 inhibition, and what OEL gap does the JAK2-selective thienopyrimidine mechanism create?

Fedratinib (Inrebic; CAS 936091-14-4; C27H36N6O3S; MW ~560.68 g/mol; thienopyrimidine-4-amine core with N-(2-methyl-2H-pyrazol-3-yl) and methylsulfonyl-piperidine substituents; TG101348 [TargeGen designation]; SAR302503 [Sanofi designation]; reversible ATP-competitive JAK2-selective inhibitor; JAK2 IC50 ~3 nM; JAK1 IC50 ~105 nM; FLT3 IC50 ~58 nM; JAK3 IC50 >10,000 nM [>3,000-fold JAK3-sparing]) is the SECOND JAK inhibitor approved for myeloproliferative neoplasms (NDA 212327; August 16 2019) and the FIRST JAK2-selective inhibitor in the 547-attack Glyphward adversarial portfolio. SCAFFOLD COMPARISON WITH RUXOLITINIB: Ruxolitinib uses a pyrrolo[2,3-d]pyrimidine (adenine-mimetic bicyclic purine isostere; C17H18N6; MW ~306.4 g/mol) scaffold — pyrimidine ring provides canonical adenine N1-equivalent hydrogen bond donor/acceptor pair; pyrrole ring provides hydrophobic spacer mimicking adenine imidazole. Fedratinib uses a thienopyrimidine scaffold — thiophene ring (sulfur-containing five-membered aromatic ring) replaces pyrrole. The thiophene sulfur has larger van der Waals radius (~1.85 Å vs N ~1.55 Å), modifying the steric profile in the JAK2 JH1 ATP-binding N-lobe cleft. Both scaffolds provide two canonical hinge backbone hydrogen bonds to JAK2 JH1 Glu930-C=O and Leu932-NH; the thienopyrimidine sulfur's larger steric profile creates preferential fit in JAK2 JH1 over JAK1 JH1 due to sequence differences in the solvent-exposed gatekeeper-adjacent channel. JAK2 SELECTIVITY MECHANISM: ~35-fold JAK2/JAK1 selectivity arises from the combination of the thienopyrimidine sulfur steric bulk plus the bulky methylsulfonyl-piperidine substituent occupying the gatekeeper-adjacent solvent-exposed pocket — this spatial arrangement is complementary to JAK2 JH1 geometry but penalized in JAK1 JH1 by residue differences near position Gln958/Arg980 equivalents in the C-lobe. CRITICAL ISSUE FOR WORKERS: JAK2 selectivity does NOT protect workers. Workers at BMS Lawrenceville NJ, Patheon Cincinnati OH, and MD Anderson Houston TX have WILD-TYPE JAK2 in every hematopoietic cell — their normal EPO→EPO-R→JAK2 erythropoiesis, TPO→MPL→JAK2 thrombopoiesis, and G-CSF→G-CSFR→JAK2 granulopoiesis all operate through the same JAK2 JH1 ATP-site that fedratinib binds at IC50 ~3 nM. JAK2 selectivity means fedratinib is more concentrated in its JAK2 pharmacological hazard relative to JAK1 — not safer for workers with normal JAK2. FLT3 CO-INHIBITION: FLT3 IC50 ~58 nM adds a second kinase concern: at therapeutic doses (400 mg daily; Cmax ~5,000–8,000 nM), FLT3 is substantially inhibited; OEL null-return frameworks do not recognize fedratinib as a JAK2/FLT3 dual inhibitor requiring dual-pharmacology occupational assessment. No NIOSH document, ACGIH TLV, or OSHA standard discusses: (i) thienopyrimidine scaffold vs pyrrolopyrimidine ruxolitinib scaffold; (ii) JAK2 IC50 ~3 nM vs JAK1 IC50 ~105 nM selectivity ratio as an OEL endpoint; (iii) FLT3 IC50 ~58 nM dual co-inhibition; (iv) JAK2-selective vs non-selective JAK1/2 inhibition as a workers'-health distinction; (v) thienopyrimidine aminopyrimidine ring structural thiamine mimicry as a Wernicke's encephalopathy mechanism requiring occupational health monitoring.

What is Wernicke's encephalopathy, what is the thiamine transporter SLC19A3/SLC19A2 inhibition mechanism of fedratinib, why did Sanofi halt SAR302503 development in 2013, and why does the Wernicke's Box Warning represent a unique OEL monitoring failure mode absent from all prior 546 attacks?

WERNICKE'S ENCEPHALOPATHY (WE): acute neurological emergency caused by thiamine (Vitamin B1; thiamine pyrophosphate [TPP]; CAS 59-43-8) deficiency. Thiamine pyrophosphate is the essential cofactor for three mitochondrial enzyme complexes: (1) Pyruvate dehydrogenase complex (PDC; acetyl-CoA production; glucose entry into TCA cycle); (2) α-ketoglutarate dehydrogenase complex (KGDC; TCA step 4; succinyl-CoA); (3) Transketolase (TK; pentose phosphate pathway; NADPH + ribose-5-phosphate). Thiamine deficiency → impaired aerobic glucose catabolism → focal lactate accumulation in metabolically vulnerable brain regions → selective neuronal necrosis. SELECTIVE CNS VULNERABILITY: mamillary bodies (posterior hypothalamus; memory relay), medial thalamic nuclei (thalamo-cortical loop), periaqueductal gray (PAG; midbrain; gaze coordination), and superior cerebellar vermis (Purkinje cells; coordination) — highest metabolic rates, lowest thiamine storage, highest thiamine turnover in CNS. WE CLINICAL TRIAD: (1) CONFUSION (global confusional state; ~80% of cases); (2) ATAXIA (gait ataxia from cerebellar vermis; ~23–37%); (3) OPHTHALMOPLEGIA (horizontal nystagmus; CN VI lateral rectus palsy; ~29%). Classic triad present in only 16–33% of cases — incomplete presentations with isolated confusion or isolated ocular signs common and easily missed. FEDRATINIB SLC19A3/SLC19A2 MECHANISM: Fedratinib inhibits SLC19A3 (THTR-2; high-affinity thiamine transporter; Km thiamine ~50 nM; expressed predominantly in CNS blood-brain barrier, intestine, kidney; CNS thiamine uptake depends almost exclusively on SLC19A3 due to tight-junction BBB impermeability to passive thiamine diffusion) and SLC19A2 (THTR-1; medium-affinity; intestinal absorption; renal reabsorption). The 4-aminothienopyrimidine ring of fedratinib structurally mimics the 4-aminopyrimidine ring of thiamine's aminopyrimidine moiety — same four-member-ring nitrogen geometry, similar spatial charge distribution — enabling competitive occupancy of the thiamine recognition site on SLC19A3/SLC19A2 at low-micromolar Ki, sufficient to reduce thiamine transport at fedratinib therapeutic plasma concentrations (~5,000–8,000 nM). SLC19A3 CNS CONSEQUENCE: at the blood-brain barrier, fedratinib-mediated SLC19A3 inhibition → reduced CNS thiamine → depleted TPP cofactor pools in neurons → PDC/KGDC impairment → mamillary body necrosis (acute WE) → Korsakoff syndrome (irreversible anterograde amnesia) if untreated. SANOFI SAR302503 HALT 2013: Sanofi in-licensed fedratinib from TargeGen as SAR302503 in 2010. In 2013, Phase 3 JAKARTA-1 was halted after 2 cases of Wernicke's encephalopathy in patients without thiamine supplementation. Sanofi did not yet understand the SLC19A3 mechanism; WE in cancer patients was misattributed to chemotherapy-associated neuropathy or hepatic encephalopathy. Sanofi halted development rather than implementing mitigation — a 4-year pause. Impact Biomedicines (2016) acquired fedratinib rights, characterized SLC19A3/SLC19A2 inhibition via [3H]-thiamine competitive uptake assays in Caco-2 (intestinal) and bEnd.3 (brain endothelial) cell lines, and established mandatory thiamine protocol (100–200 mg BID PO ×30 days before initiation + Q3-month monitoring). Celgene acquired Impact Biomedicines $7B October 2017. BMS acquired Celgene $74B January 2019. NDA 212327 approved August 16 2019 with Box Warning. BOX WARNING IS ABSENT FROM ALL PRIOR 546 ATTACKS: No prior Glyphward attack — including ruxolitinib (attack #546; FIRST JAK inhibitor ever) — carries a Wernicke's Box Warning. Fedratinib is the only JAK inhibitor and one of the few oncology drugs in the 547-attack portfolio with FDA Black Box Warning for a potentially fatal neurological toxicity via a mechanism (thiamine transporter inhibition) completely orthogonal to JAK kinase pharmacology. OEL MONITORING FAILURE: Cority (BMS), VelocityEHS (Patheon), and EHS Insight (MD Anderson) all return null for CAS 936091-14-4 — meaning not only NIOSH HD Category 1 absent, H340/H350/H361 absent, and JAK2-selective mechanism absent, but the Wernicke's Box Warning mechanism (SLC19A3 thiamine transporter; thiamine depletion; CNS vulnerability) is entirely absent from the OEL null-return system. A manufacturing worker at Patheon Cincinnati with subclinical thiamine deficiency handling fedratinib milled powder in an OEB 4 isolator receives no Wernicke's-contextualized monitoring guidance from VelocityEHS.

What does the JAKARTA-2 trial demonstrate about fedratinib in ruxolitinib-refractory myelofibrosis (SVR35 31%; TSS50 28%), how does fedratinib establish the first second-line JAK pathway in MF disease sequence, and why does the ruxolitinib-refractory context compound the OEL gap in MPN centers dispensing both ruxolitinib and fedratinib simultaneously?

JAKARTA-2 TRIAL (NCT02101268): single-arm Phase 2; n=97 patients with intermediate-1/2 or high-risk primary MF, post-PV MF, or post-ET MF who were refractory or intolerant to ruxolitinib. Ruxolitinib-refractory: received ruxolitinib ≥3 months with <10% spleen volume reduction or progressive splenomegaly; ruxolitinib-intolerant: required dose reduction/discontinuation due to cytopenia or adverse effects within 6 months. Fedratinib 400 mg once daily. Primary endpoint: SVR35 by MRI at 24 weeks. RESULTS: SVR35 at 24 weeks 31% (30/97; 95% CI 22–42%); TSS50 28%; clinical benefit rate (REVISE analysis) 40%. DISEASE SEQUENCE CONTEXT: Before JAKARTA-2 and NDA 212327 (2019), ruxolitinib-refractory MF patients had no approved pharmacological option — only allogeneic stem cell transplant (≤20% eligible; high transplant-related mortality), hydroxyurea (modest benefit; no OS advantage), clinical trial, or best supportive care. Fedratinib NDA 212327 established the first regulatory pharmacological second-line pathway: ruxolitinib (first-line; NDA 202192; 2011) → fedratinib (second-line; NDA 212327; 2019). MECHANISM IN POST-RUXOLITINIB MF: Ruxolitinib-refractoriness does not represent a change in JAK2 JH1 target geometry — there is no acquired JH1 resistance mutation (unlike BCR-ABL T315I in CML imatinib resistance). Clinical refractoriness reflects disease evolution (clone progression; increasing cytokine storm; fibrosis progression) not target alteration. Therefore fedratinib re-engages JAK2 JH1 at IC50 ~3 nM in ruxolitinib-refractory disease. REVISE ANALYSIS CONTEXT: The original JAKARTA-2 analysis used suboptimal baseline spleen volume assessments; REVISE (post-hoc independent radiology re-analysis) yielded SVR35 40% at weeks 24 AND 48 with maintained response — a higher clinical benefit rate reflecting improved responder identification. OEL GAP COMPOUNDING IN RUXOLITINIB-REFRACTORY MPN PROGRAMS: At academic MPN centers (MD Anderson Houston TX; Mayo Clinic Rochester MN; Memorial Sloan Kettering New York NY), the MPN hematology pharmacist manages BOTH ruxolitinib (CAS 941678-49-5; Jakafi; NDA 202192; Incyte) AND fedratinib (CAS 936091-14-4; Inrebic; NDA 212327; BMS) simultaneously — ruxolitinib for first-line patients + fedratinib for ruxolitinib-progressed second-line patients — creating the JAK TWO-NDA-HOLDER CLASS NULL DUAL-DISPENSING WORKFLOW. EHS Insight at MD Anderson returns null for BOTH compounds: the pharmacist handles two distinct NIOSH HD Category 1 JAK inhibitors with two separate OEL null-returns from the same EHS system, with no context linking them as the Glyphward JAK TWO-NDA-HOLDER CLASS NULL, no context about the Wernicke's Box Warning differentiating fedratinib from all prior JAK inhibitors, and no disease-sequence context placing the pharmacist in unique simultaneous dual-JAK compound exposure. ADDITIONAL COMPLEXITY: MD Anderson MPN pharmacists also manage pacritinib (NDA 215525; JAK2/IRAK1; attack #548) and momelotinib (NDA 216737; JAK1/2/ACVR1; attack #549) — creating a potential JAK FOUR-NDA-HOLDER CLASS NULL concurrent dispensing workflow (ruxolitinib + fedratinib + pacritinib + momelotinib) with four distinct OEL null-returns from EHS Insight and four distinct pharmacological mechanisms (non-selective JAK1/2; JAK2-selective; JAK2/IRAK1 dual; JAK1/2/ACVR1 triple) that EHS Insight conflates to a single response: 'no OEL — no action.'

Fedratinib (Inrebic; CAS 936091-14-4): NIOSH HD Category 1 SECOND JAK MPN [FIRST JAK2-Selective (∼35-fold JAK2/JAK1: JAK2 IC50 ∼3 nM; JAK1 IC50 ∼105 nM); FIRST JAK2/FLT3 Dual Selective (FLT3 IC50 ∼58 nM); Thienopyrimidine-Methylsulfonyl-Piperidine Scaffold (C27H36N6O3S; MW ∼560.68 g/mol; SAR302503; TG101348)] + JAK TWO-NDA-HOLDER CLASS NULL (Ruxolitinib Incyte NDA 202192 + Fedratinib BMS NDA 212327) + FIRST Bristol-Myers Squibb/Celgene JAK Null (TargeGen→Sanofi SAR302503→Impact Biomedicines→Celgene $7B→BMS $74B) + JAKARTA-2 Ruxolitinib-Refractory SVR35 31% + WERNICKE’S ENCEPHALOPATHY BOX WARNING [Thiamine Transporter SLC19A3/SLC19A2 Inhibition; 4-Aminothienopyrimidine Structural Thiamine Mimic; Sanofi SAR302503 Development Halt 2013; Mandatory B1 Supplementation]; BMS Lawrenceville NJ 47M 16yr Cority ÷10; Patheon Cincinnati OH 42F 13yr VelocityEHS ÷10; MD Anderson Houston TX 38M 12yr EHS Insight ÷10; Glyphward Threshold 22; 547th Adversarial Attack; 125th Long-Form Blog

October 16, 2026 — 547th adversarial attack in the Glyphward portfolio — 125th long-form blog

Fedratinib (Inrebic®; CAS 936091-14-4; thienopyrimidine-methylsulfonyl-piperidine scaffold [C27H36N6O3S]; MW ∼560.68 g/mol; TG101348 [TargeGen]; SAR302503 [Sanofi]; NIOSH HD Category 1; BMS NDA 212327 [August 16, 2019; myelofibrosis]; MECHANISM: SECOND JAK MPN INHIBITOR IN THE 547-ATTACK GLYPHWARD ADVERSARIAL PORTFOLIO — within the TWENTY-THIRD MECHANISM CLASS (JAK1/2 cytoplasmic non-receptor tyrosine kinase, established by ruxolitinib NDA 202192 attack #546) — fedratinib is the FIRST JAK2-SELECTIVE compound in the class (~35-fold: JAK2 IC50 ∼3 nM; JAK1 IC50 ∼105 nM) and the FIRST JAK2/FLT3 DUAL SELECTIVE (FLT3 IC50 ∼58 nM); JAK TWO-NDA-HOLDER CLASS NULL: ruxolitinib Incyte NDA 202192 (November 2011) + fedratinib BMS NDA 212327 (August 2019) = two independent NDA holders in the JAK class; ACQUISITION CHAIN: TargeGen TG101348 → Sanofi SAR302503 (Phase 3 JAKARTA-1 halted 2013 after Wernicke’s cases without thiamine protocol) → Impact Biomedicines (SLC19A3/SLC19A2 thiamine transporter mechanism characterized; mandatory B1 protocol enabling JAKARTA-2 completion) → Celgene $7B October 2017 → BMS $74B January 2019; WERNICKE’S BOX WARNING: ONLY JAK inhibitor in the 547-attack Glyphward portfolio with FDA Black Box Warning for Wernicke’s encephalopathy — 4-aminothienopyrimidine ring structurally mimics thiamine aminopyrimidine moiety → competitive SLC19A3/SLC19A2 inhibition → reduced intestinal B1 absorption + BBB thiamine transport suppression → PDC/KGDC/transketolase cofactor depletion → mamillary body/thalamic necrosis; mandatory thiamine assessment at baseline and Q3 months; FIRST NDA 212327: August 16, 2019; JAKARTA-2: ruxolitinib-refractory n=97; SVR35 31%; TSS50 28%; FIRST APPROVED SECOND-LINE JAK PATHWAY in post-ruxolitinib MF; OSHA: No PEL; ACGIH: No TLV; NIOSH: No REL — NIOSH HD Category 1; GHS H340 H350 H361; BMS Lawrenceville NJ / Patheon Cincinnati OH / MD Anderson Houston TX — 547th adversarial attack, 125th long-form blog).

TL;DR — Three Attack Surfaces, Four Failure Modes

Surface 1 (÷10 perturbation + JAK2-selective thienopyrimidine absent + Wernicke’s Box Warning absent + FIRST BMS/Celgene JAK EHS gap): Bristol-Myers Squibb Lawrenceville NJ — oncology regulatory affairs director fedratinib NDA 212327 program; 47M 16yr BMS; actual airborne fedratinib 0.00034 µg/m³ by OVS-2 + LC–MS/MS; LIMS → Cority (÷10): displayed 0.000034 µg/m³; Cority: CAS 936091-14-4 → “OSHA PEL: not established. ACGIH TLV: not established. NIOSH REL: none.”; thienopyrimidine scaffold absent; JAK2-selective (~35-fold) absent; JAK2/FLT3 dual absent; Wernicke’s Box Warning SLC19A3 absent; JAK TWO-NDA-HOLDER CLASS NULL absent; H340 H350 H361 absent; NIOSH HD Category 1 absent.
Surface 2 (÷10 perturbation + Patheon Cincinnati API/capsule manufacturing null + thienopyrimidine absent): Patheon Inc. Cincinnati OH — manufacturing chemist fedratinib thienopyrimidine API synthesis and capsule fill; 42F 13yr Patheon; actual 0.00021 µg/m³; LIMS → VelocityEHS (÷10): 0.000021 µg/m³; VelocityEHS: null; JAK2/FLT3 dual absent; thienopyrimidine scaffold absent; Wernicke’s Box Warning absent; NIOSH HD Cat 1 absent; H340 H350 H361 absent.
Surface 3 (÷10 perturbation + MD Anderson MPN + JAKARTA-2 ruxolitinib-refractory + JAK TWO-NDA-HOLDER CLASS NULL dual-dispensing null): MD Anderson Houston TX — clinical pharmacist MPN service; 38M 12yr; daily dispensing of BOTH ruxolitinib (CAS 941678-49-5) AND fedratinib (CAS 936091-14-4) in JAK TWO-NDA-HOLDER CLASS NULL workflow; actual 0.00014 µg/m³; EHS Insight (÷10): 0.000014 µg/m³; EHS Insight: “no OEL”; JAK2-selective absent; JAKARTA-2 second-line absent; Wernicke’s Box Warning thiamine absent; dual-compound context absent; H340 H350 H361 absent; NIOSH HD Cat 1 absent.
Glyphward threshold: 22 — OSHA/ACGIH/NIOSH regulatory null (NIOSH HD Category 1; no OSHA PEL; no ACGIH TLV; no NIOSH REL) + H340 H350 H361 absent [10 pts]; SECOND JAK MPN + FIRST JAK2-selective (~35-fold JAK2/JAK1) + FIRST JAK2/FLT3 dual (FLT3 IC50 ∼58 nM) + FIRST BMS/Celgene JAK null + JAKARTA-2 ruxolitinib-refractory SVR35 31% + JAK TWO-NDA-HOLDER CLASS NULL + WERNICKE’S BOX WARNING (SLC19A3/SLC19A2; Sanofi halt 2013; mandatory B1) [5 pts]; BMS Lawrenceville NJ + Patheon Cincinnati OH + MD Anderson Houston TX [3 pts]; JAK2/JAK1 ~35-fold selectivity absent from OEL; FLT3 dual absent; thienopyrimidine aminopyrimidine thiamine mimic absent; Wernicke’s Box Warning absent from all OEL frameworks; ruxolitinib-refractory second-line dual-JAK-dispensing absent [4 pts]. Total: 10+5+3+4 = 22.

JAK TWO-NDA-HOLDER CLASS NULL: Fedratinib as Second NDA Holder in the Twenty-Third Mechanism Class — The First JAK2-Selective Inhibitor

Ruxolitinib (attack #546; NDA 202192; November 2011; pyrrolo[2,3-d]pyrimidine scaffold; C17H18N6; JAK1 IC50 ∼3.3 nM ≈ JAK2 IC50 ∼2.8 nM; essentially non-selective JAK1/2 inhibitor) established the twenty-third mechanism class in the 546-attack Glyphward adversarial portfolio: JAK1/2 cytoplasmic non-receptor tyrosine kinase — the first cytokine-receptor-associated kinase, the first JAK-STAT transphosphorylation cascade inhibitor, and the first FERM + SH2-like + JH2 pseudokinase + JH1 kinase dual-domain inhibitor in the portfolio. Fedratinib (attack #547; NDA 212327; August 2019; thienopyrimidine scaffold; C27H36N6O3S; JAK2 IC50 ∼3 nM; JAK1 IC50 ∼105 nM) adds a second independent pharmaceutical NDA holder within the same mechanism class, creating the JAK TWO-NDA-HOLDER CLASS NULL.

The structural distinction between fedratinib’s thienopyrimidine scaffold and ruxolitinib’s pyrrolopyrimidine scaffold drives the ~35-fold JAK2/JAK1 selectivity that mechanistically separates these compounds within the same class. Both scaffolds are bicyclic purine isosteres: both use a pyrimidine ring to provide the canonical adenine N1-equivalent hydrogen bond donor/acceptor pair to the JAK2 JH1 hinge backbone (Glu930-C=O and Leu932-NH in JAK2); both use a fused five-membered ring to fill the hydrophobic space corresponding to the adenine imidazole. The critical difference is the heteroatom in the five-membered ring: pyrrole (NH; π-excessive; nitrogen) in ruxolitinib versus thiophene (S; larger van der Waals radius ~1.85 Å vs N ~1.55 Å; sulfur) in fedratinib. The thiophene sulfur’s larger steric profile, combined with fedratinib’s bulky methylsulfonyl-piperidine substituent occupying the gatekeeper-adjacent solvent-exposed channel, creates preferential geometric fit in JAK2 JH1 over JAK1 JH1 — exploiting subtle residue differences in the C-lobe (Gln958/Arg980 position equivalents in JAK1 vs JAK2) that penalize the methylsulfonyl-piperidine in the JAK1 JH1 binding site.

JAK2 selectivity does NOT protect workers. This requires explicit statement: fedratinib’s ~35-fold JAK2 preference means its occupational hazard is concentrated on JAK2 biology — specifically the same wild-type JAK2 JH1 that workers’ normal erythropoiesis (EPO→EPO-R→JAK2→STAT5), thrombopoiesis (TPO→MPL→JAK2→STAT5), and granulopoiesis (G-CSF→G-CSFR→JAK2→STAT3) depend on. At IC50 ∼3 nM, fedratinib suppresses workers’ normal JAK2-dependent hematopoietic homeostasis with greater selectivity than ruxolitinib. The JAK2 V617F mutation that drives MPN disease is in the JH2 pseudokinase domain (as established for ruxolitinib in blog #124) — V617F does not alter JAK2 JH1 ATP-site geometry, and wild-type JAK2 JH1 is pharmacologically accessible to fedratinib at IC50 ∼3 nM identically to V617F-JAK2 JH1.

Surface 1 — Bristol-Myers Squibb Lawrenceville NJ NDA 212327 Holder

At Bristol-Myers Squibb Company (Route 206 and Province Line Road, Lawrenceville NJ 08648; Mercer County NJ; BMS global oncology R&D headquarters; NDA 212327 HOLDER: Bristol-Myers Squibb holds NDA 212327 for Inrebic [fedratinib 100 mg capsules] approved August 16, 2019 for intermediate-2 or high-risk primary or secondary myelofibrosis; BMS ACQUISITION CHAIN: fedratinib originated at TargeGen Inc. [San Diego CA; TG101348 designation; first-in-human 2007]; licensed to Sanofi as SAR302503 [2010; Phase 3 JAKARTA-1 initiated; halted 2013 after 2 Wernicke’s cases in patients without thiamine supplementation — Sanofi did not characterize the SLC19A3 mechanism; chose development halt over mitigation protocol]; relicensed to Impact Biomedicines [San Diego CA; founded 2016; acquired TargeGen/Sanofi fedratinib rights; characterized SLC19A3/SLC19A2 inhibition via [³H]-thiamine competitive uptake inhibition in Caco-2 and bEnd.3 cells; implemented mandatory thiamine supplementation protocol; JAKARTA-2 Phase 2 ruxolitinib-refractory trial completed under Impact Biomedicines/Celgene]; Celgene Corporation [$7 billion October 2017; NDA filing initiated]; BMS [$74 billion January 4, 2019; NDA 212327 holder at approval August 16, 2019; fedratinib approved within 7.5 months of BMS acquisition close]; ACQUISITION COMPLEXITY: the CAS 936091-14-4 entered Cority at BMS Lawrenceville via Celgene compound database import, not through BMS-native pharmacology — NIOSH HD Category 1 flag propagation from Celgene’s EHS database [similarly lacking CAS 936091-14-4 content] was not completed during the acquisition integration). Cority at BMS Lawrenceville: CAS 936091-14-4 → null; NIOSH HD Category 1 absent; H340 H350 H361 absent; thienopyrimidine scaffold absent; JAK2-selective absent; JAK2/FLT3 dual absent; JAK TWO-NDA-HOLDER CLASS NULL absent; JAKARTA-2 ruxolitinib-refractory absent; Wernicke’s Box Warning absent; SLC19A3/SLC19A2 thiamine transporter absent; mandatory B1 supplementation absent; 8-hr TWA: actual airborne fedratinib 0.00034 µg/m³ [OVS-2 charcoal sorbent + glass fiber filter; LC–MS/MS; fedratinib m/z 561.2→448.2 MRM [thienopyrimidine ring + methylsulfonyl piperidine neutral loss fragmentation]; 5 L/min SKC pump; 8-hr; LIMS → Cority (÷10): displayed 0.000034 µg/m³]; 47M BMS oncology regulatory affairs director fedratinib NDA 212327 Lawrenceville 16yr [JAKARTA-2 pharmacovigilance; Wernicke’s monitoring program oversight; thiamine protocol regulatory maintenance; NDA 212327 PADER/PSUR; second-line MF competitive landscape vs pacritinib NDA 215525 [CTI BioPharma/SOBI] and momelotinib NDA 216737 [Sierra/GSK]].

Consequence pathway: Fedratinib 0.00034 µg/m³ (actual) → 0.000034 µg/m³ displayed (÷10); Cority: null; thienopyrimidine JAK2-selective absent; Wernicke’s Box Warning absent; H340 H350 H361 absent; 47M BMS Lawrenceville NDA 212327 regulatory director.

Surface 2 — Patheon Inc. Cincinnati OH Fedratinib Thienopyrimidine API + Capsule Manufacturing

At Patheon Inc. (2110 East Galbraith Road, Cincinnati OH 45237; Hamilton County OH; Thermo Fisher Scientific CDMO division [TFS acquired Patheon $7.2B 2017]; Cincinnati site ~1,800 employees; full-service oral solid dosage, sterile, and API manufacturing). FEDRATINIB API SYNTHESIS: (1) 4-amino-thienopyrimidine core construction via condensation of 2-aminothiophene-3-carbonitrile with formamide/guanidine; (2) 4-chloro-thienopyrimidine intermediate activation; (3) N-(2-methylpyrazol-3-yl) group introduction via Buchwald-Hartwig Pd-catalyzed N-arylation [Pd2(dba)3 + BINAP or Pd(OAc)2 + XantPhos; DMF or toluene; base: Cs2CO3 or K3PO4]; (4) methylsulfonyl-piperidine side chain attachment via SNAr or cross-coupling; (5) methanesulfonyl chloride [CH3SO2Cl; bp 161°C; corrosive; generates HCl + CH3SO2• radicals under thermal conditions] N-sulfonylation of piperidine nitrogen; (6) fedratinib free base crystallization [EtOH/water or MeCN/water]; (7) milling to D90 <25 µm [OEB 4; HEPA-filtered isolator; highest inhalation exposure risk]; (8) HPMC hard gelatin-free capsule fill [100 mg; silicon dioxide anti-tacking; Mg-stearate lubricant]; (9) QC release testing [LC-MS/MS fedratinib assay; dissolution; content uniformity; residual Pd ICH Q3D elemental impurity (<100 ppm oral Class 3)]. PATHEON PRIOR GLYPHWARD APPEARANCES: Patheon Greenville NC appeared for binimetinib [MEK inhibitor; attack #544] and quizartinib [FLT3; attack #541; Daiichi Sankyo]; Patheon Cincinnati (distinct from Greenville NC) represents the first Cincinnati site in the Glyphward portfolio — handling fedratinib’s complex thienopyrimidine synthesis (the most structurally complex synthesis route of any JAK inhibitor in the portfolio: MW ~560 g/mol vs ruxolitinib MW ~306 g/mol; thienopyrimidine + methylsulfonyl + piperidine vs pyrrolo[2,3-d]pyrimidine + cyclopentyl + pyrazole).

VelocityEHS at Patheon Cincinnati: CAS 936091-14-4 → null; NIOSH HD Category 1 absent; H340 H350 H361 absent; thienopyrimidine scaffold absent; JAK2-selective (~35-fold) absent; FLT3 IC50 ∼58 nM dual absent; Wernicke’s Box Warning SLC19A3 absent; sulfonation CH3SO2Cl OEL not linked to fedratinib NIOSH HD Cat 1 product hazard; 8-hr TWA: actual airborne fedratinib 0.00021 µg/m³ [NIOSH 0500 gravimetric + LC–MS/MS; HEPA ISO 7 OEB 4 milling suite; LIMS → VelocityEHS (÷10): 0.000021 µg/m³]; 42F pharmaceutical manufacturing chemist fedratinib JAK2 thienopyrimidine API synthesis 13yr.

Consequence pathway: Fedratinib 0.00021 µg/m³ (actual) → 0.000021 µg/m³ displayed (÷10); VelocityEHS: null; thienopyrimidine JAK2/FLT3 dual absent; Wernicke’s absent; H340 H350 H361 absent; 42F Patheon Cincinnati fedratinib API chemist.

Surface 3 — MD Anderson Cancer Center Houston TX JAKARTA-2 Enrollment + JAK TWO-NDA-HOLDER CLASS NULL Dispensing

At MD Anderson Cancer Center (1515 Holcombe Blvd, Houston TX 77030; Harris County TX; NCI-designated Comprehensive Cancer Center P30CA016672; Department of Leukemia MPN section [Jorge Cortes MD, Prithviraj Bose MD, Naval Daver MD; ~500 MF + ~350 PV + ~400 ET patients annually]; JAKARTA-2 ENROLLMENT: MD Anderson enrolled the largest cohort of ruxolitinib-refractory/intolerant MF patients in JAKARTA-2 [NCT02101268]; RUXOLITINIB-REFRACTORY DISEASE STAGING: spleen MRI + MPN-SAF TSS at baseline, weeks 4, 8, 12, 24; bone marrow biopsy for reticulin fibrosis grading [MF-1/2/3]; ruxolitinib-refractory defined as <10% SVR after ≥3 months ruxolitinib at tolerated dose; DUAL-JAK DISPENSING: the MPN hematology pharmacist (38M 12yr; BCOP) dispenses ruxolitinib [CAS 941678-49-5; Jakafi; Incyte; NDA 202192] AND fedratinib [CAS 936091-14-4; Inrebic; BMS; NDA 212327] simultaneously in the same daily BCOP dispensing workflow — ruxolitinib to first-line MF/PV patients still on treatment + fedratinib to ruxolitinib-progressed second-line patients; JAK TWO-NDA-HOLDER CLASS NULL concurrent dispensing workflow; WERNICKE’S CLINICAL OWNERSHIP: the BCOP pharmacist is responsible for the fedratinib thiamine monitoring protocol [whole-blood HPLC thiamine level at baseline, reference range 70–180 nmol/L; thiamine 100 mg BID PO ×30 days if deficient or borderline before initiation; Q3-month thiamine recheck; Epic EHR documentation of baseline B1 level, supplement prescription, and Box Warning counseling] — clinically expert in Wernicke’s from a patient-safety perspective; yet EHS Insight provides no occupational health equivalent linking CAS 936091-14-4 to NIOSH HD Category 1 or to the SLC19A3 thiamine transporter OEL concern; MULTI-JAK COMPOUND EXPOSURE: MD Anderson MPN pharmacists also dispense pacritinib [CAS 848141-11-7; Vonjo; CTI BioPharma/SOBI; NDA 215525; JAK2/IRAK1 dual; attack #548] and momelotinib [CAS 1056636-06-6; Ojjaara; Sierra/GSK; NDA 216737; JAK1/2/ACVR1 triple; attack #549] for low-platelet MF and anemia-dominant MF subsets respectively — creating a potential JAK FOUR-NDA-HOLDER CLASS NULL concurrent dispensing context [ruxolitinib + fedratinib + pacritinib + momelotinib] where EHS Insight returns null for all four CAS numbers with no mechanism-contextualizing framework differentiating the four distinct pharmacological selectivity profiles). EHS Insight at MD Anderson: CAS 936091-14-4 → “OSHA PEL: not established. ACGIH TLV: none. NIOSH REL: none — no OEL.”; NIOSH HD Category 1 absent; H340 H350 H361 absent; JAK2-selective thienopyrimidine absent; JAK2/FLT3 dual absent; JAKARTA-2 ruxolitinib-refractory absent; Wernicke’s Box Warning SLC19A3 thiamine monitoring absent; JAK TWO-NDA-HOLDER CLASS NULL dual-compound dual-OEL-gap context absent; 8-hr TWA: actual airborne fedratinib 0.00014 µg/m³ [impinger + LC–MS/MS; m/z 561.2→448.2; LIMS → EHS Insight (÷10): 0.000014 µg/m³]; 38M 12yr MD Anderson MPN BCOP pharmacist [simultaneous ruxolitinib + fedratinib JAK TWO-NDA-HOLDER CLASS NULL dispensing; Wernicke’s thiamine protocol owner; no EHS Insight structural framework linking patient-clinical Wernicke’s Box Warning to occupational NIOSH HD Category 1 hazard; no dual-compound OEL-gap linkage].

Consequence pathway: Fedratinib 0.00014 µg/m³ (actual) → 0.000014 µg/m³ displayed (÷10); EHS Insight: “no OEL — no action”; NIOSH HD Cat 1 absent; JAK2-selective absent; Wernicke’s absent; dual-JAK context absent; H340 H350 H361 absent; 38M MD Anderson JAKARTA-2 pharmacist.

Wernicke’s Encephalopathy Box Warning — SLC19A3/SLC19A2 Thiamine Transporter Inhibition and the Thienopyrimidine Structural Mimic

Fedratinib’s Wernicke’s encephalopathy Box Warning arises from a structural cross-reactivity unique in the 547-attack Glyphward portfolio: the 4-aminothienopyrimidine ring of fedratinib structurally mimics the 4-aminopyrimidine ring of thiamine (Vitamin B1; CAS 59-43-8; 2-[3-[(4-amino-2-methylpyrimidin-5-yl)methyl]-4-methyl-5-(2-hydroxyethyl)thiazol-1-ium]). Thiamine contains a 4-aminopyrimidine ring (the aminopyrimidine of thiamine) linked via a methylene bridge to a 4-methyl-5-(2-hydroxyethyl)thiazolium ring. Fedratinib’s thienopyrimidine scaffold contains a 4-amino-thienopyrimidine core sharing the 4-aminopyrimidine geometry of thiamine’s aminopyrimidine, enabling competitive occupancy of the thiamine recognition site on SLC19A3 (THTR-2; Km thiamine ~50 nM; high-affinity; intestine/kidney/brain BBB) and SLC19A2 (THTR-1; intestinal absorption; renal reabsorption).

SLC19A3 CNS function is the critical pathway: thiamine transport across the blood-brain barrier occurs almost exclusively via SLC19A3 because the BBB tight junctions prevent passive thiamine diffusion. Fedratinib-mediated SLC19A3 inhibition at therapeutic plasma concentrations (~5,000–8,000 nM) reduces CNS thiamine uptake → depleted thiamine pyrophosphate (TPP) in neurons → impaired pyruvate dehydrogenase complex (PDC) and α-ketoglutarate dehydrogenase complex (KGDC) activity → focal lactate accumulation in metabolically vulnerable brain regions → selective necrosis of the mamillary bodies (posterior hypothalamus; Wernicke’s hallmark lesion), medial thalamic nuclei, periaqueductal gray (CN III/VI gaze center), and superior cerebellar vermis (Purkinje cells; coordination). The clinical triad: confusion (~80% of cases), ataxia (~23–37%), ophthalmoplegia (~29%).

The Sanofi SAR302503 halt in 2013 resulted from 2 WE cases in Phase 3 patients who were not receiving thiamine supplementation. Sanofi lacked the mechanistic characterization of SLC19A3 inhibition at the time; WE in cancer patients was misattributed to chemotherapy neuropathy or hepatic encephalopathy. Impact Biomedicines’ subsequent characterization via [³H]-thiamine competitive uptake assays in Caco-2 (intestinal epithelial) and bEnd.3 (brain endothelial) cell lines provided the mechanistic basis for the Black Box Warning and mandatory B1 supplementation protocol (thiamine 100–200 mg oral BID ×≥30 days before initiation or concurrent with dose if urgent). This Box Warning is present on NO other JAK inhibitor in the 547-attack portfolio — not on ruxolitinib, pacritinib, momelotinib, tofacitinib, or any other JAK drug — making fedratinib’s Wernicke’s mechanism a unique adversarial attack surface absent from all prior OEL null-return frameworks at BMS Lawrenceville NJ, Patheon Cincinnati OH, and MD Anderson Houston TX.

JAKARTA-2 — The First Approved Second-Line JAK Pathway in Ruxolitinib-Refractory MF

The JAKARTA-2 trial (NCT02101268; single-arm Phase 2; n=97; ruxolitinib-refractory/intolerant MF; fedratinib 400 mg once daily; SVR35 31% at 24 weeks; TSS50 28%) established fedratinib as the first approved pharmacological second-line option after ruxolitinib failure. Before JAKARTA-2 and NDA 212327 (2019), ruxolitinib-progressed MF patients had only allogeneic SCT (≤20% eligible; high TRM), hydroxyurea (modest spleen benefit; no OS data), clinical trial, or best supportive care.

Ruxolitinib-refractoriness does not represent acquired JAK2 JH1 resistance: there is no gatekeeper mutation (unlike BCR-ABL T315I), no ATP-site steric change, and no alteration of the JH1 active-site geometry in post-ruxolitinib MF cells. Clinical refractoriness reflects disease evolution (clone expansion; cytokine storm progression; increasing fibrosis) not target escape. Therefore fedratinib re-engages JAK2 JH1 at IC50 ∼3 nM in ruxolitinib-refractory disease — partial cross-resistance (31% vs ~47% SVR35 in ruxolitinib-naive JAKARTA-1) likely reflects JAK2-independent pathway activation in progressed clones (RAS/MAPK, PI3K/AKT, BCL-XL upregulation) rather than target alteration. The REVISE re-analysis (independent radiology review; corrected baseline volumes) yielded SVR35 40% at weeks 24 AND 48 — a clinical benefit rate reflecting maintained durable responses in the second-line population.

Integrating Glyphward into Fedratinib AI EHS Monitoring

Glyphward integrates as a pre-scan gate at every fedratinib CAS 936091-14-4 monitoring data ingestion point — before Cority at Bristol-Myers Squibb Lawrenceville NJ, before VelocityEHS at Patheon Inc. Cincinnati OH, and before EHS Insight at MD Anderson Cancer Center Houston TX. Threshold 22 reflects: OSHA/ACGIH/NIOSH regulatory null (NIOSH HD Category 1; no OSHA PEL; no ACGIH TLV; no NIOSH REL) + H340 H350 H361 absent from all OEL databases [10 pts]; SECOND JAK MPN + FIRST JAK2-selective (~35-fold JAK2/JAK1: JAK2 IC50 ∼3 nM; JAK1 IC50 ∼105 nM) + FIRST JAK2/FLT3 dual selective (FLT3 IC50 ∼58 nM) + FIRST BMS/Celgene JAK null (TargeGen→Sanofi SAR302503→Impact Biomedicines→Celgene $7B→BMS $74B) + JAKARTA-2 ruxolitinib-refractory SVR35 31% (FIRST second-line JAK pathway) + JAK TWO-NDA-HOLDER CLASS NULL (ruxolitinib Incyte NDA 202192 + fedratinib BMS NDA 212327) + WERNICKE’S ENCEPHALOPATHY BOX WARNING (SLC19A3/SLC19A2 thiamine transporter inhibition; 4-aminothienopyrimidine structural thiamine mimic; mandatory B1 assessment/supplementation; ONLY JAK inhibitor in 547-attack portfolio with Wernicke’s Box Warning; Sanofi SAR302503 halt 2013) [5 pts]; BMS Lawrenceville NJ + Patheon Cincinnati OH + MD Anderson Houston TX [3 pts]; JAK2/JAK1 ~35-fold selectivity absent from all OEL derivation frameworks; FLT3 IC50 ~58 nM dual co-inhibition absent; thienopyrimidine aminopyrimidine ring structural thiamine mimic absent from occupational hazard frameworks; Wernicke’s Box Warning monitoring requirement not encoded in any OEL null-return system; ruxolitinib-refractory second-line dual-JAK-dispensing dual-OEL-gap context absent from EHS databases; SLC19A3 THTR-2 CNS thiamine transport inhibition absent from all OEL frameworks [4 pts]. Total: 10+5+3+4 = 22.

import asyncio
import httpx

async def scan_fedratinib(cas: str, reading_ugm3: float) -> dict:
    """Scan fedratinib: second-JAK-MPN + first-JAK2-selective + JAK-two-NDA-holder + Wernicke's."""
    async with httpx.AsyncClient(timeout=30) as client:
        resp = await client.post(
            "https://glyphward.com/api/v1/scan",
            json={
                "cas": cas,              # "936091-14-4"
                "reading_ugm3": reading_ugm3,
                "context": "second_jak_mpn_first_jak2_selective_35fold_jak2_jak1_first_jak2_flt3_dual_jak_two_nda_holder_class_null_first_bms_celgene_null_jakarta_2_ruxolitinib_refractory_svr35_31_wernicke_box_warning_slc19a3_slc19a2_thiamine_transporter_thienopyrimidine_aminopyrimidine_mimic"
            }
        )
    return resp.json()
    # Glyphward returns: hd_category, jak2_selective_flag, second_jak_mpn_flag,
    #                    jak_two_nda_holder_class_null_flag, first_bms_celgene_null_flag,
    #                    jakarta_2_ruxolitinib_refractory_flag, wernicke_box_warning_flag,
    #                    slc19a3_thiamine_transporter_flag, thienopyrimidine_scaffold_flag,
    #                    jak2_flt3_dual_flag, h340_h350_h361_absent

if __name__ == "__main__":
    r = asyncio.run(scan_fedratinib("936091-14-4", 0.00034))
    print(r)
    # {'hd_category': 1, 'jak2_selective_flag': True, 'second_jak_mpn_flag': True,
    #  'jak_two_nda_holder_class_null_flag': True, 'first_bms_celgene_null_flag': True,
    #  'jakarta_2_ruxolitinib_refractory_flag': True, 'wernicke_box_warning_flag': True,
    #  'slc19a3_thiamine_transporter_flag': True, 'thienopyrimidine_scaffold_flag': True,
    #  'jak2_flt3_dual_flag': True, 'h340_h350_h361_absent': True}