BioSignal

GSK paid $110 million in cash for a molecule that has never been given to a human being. Two days earlier Roche paid $75 million for another one. Both sellers are Chinese, and neither asset has a clinical trial behind it. On September 3 HUTCHMED licensed HMPL-A830 to GSK. It is an anti-EGFR antibody carrying a selective KRAS small-molecule inhibitor as its payload, which the company calls a first-in-class KRAS-EGFR antibody-targeted therapy conjugate. GSK takes worldwide rights outside mainland China, Hong Kong, Macau and Taiwan, for $110 million upfront and up to $1.185 billion in milestones, plus tiered royalties. HUTCHMED runs the global Phase 1 itself, starting in the second half of this year under NCT07718581. GSK also took a right of first negotiation on one earlier-stage conjugate from the same platform. I owe a correction on this. On Thursday I closed a post about KRAS G12D by calling Tyligand's preclinical EGFR-KRAS conjugate the one thing in its pipeline that nobody ahead of it was running. GSK had paid $110 million for that architecture seven days before I wrote it. The sentence was narrowly true about the KRAS G12D companies and wrong about the field. The architecture deserves the attention. An antibody-drug conjugate delivers a cytotoxin and kills whatever it reaches. This delivers a signalling inhibitor instead, so the antibody does two jobs at once: it blocks EGFR, and it concentrates a KRAS drug inside cells that display the receptor. Every KRAS inhibitor outside G12C has run into the same wall, which is that the concentration the tumour needs is more than the rest of the patient will tolerate. HUTCHMED frames the gap as non-G12C disease and has not said which mutations the payload covers. That is the number to ask for before the Phase 1 reads out. Two days before that, Roche licensed SIM0660 from Simcere Zaiming, the oncology arm of Simcere Pharmaceutical. It is a trispecific antibody that grips CD79a and CD19 on B cells and CD3 on T cells, aimed at patients who have already failed CD19 or CD20 therapy, and at B-cell autoimmune disease. Seventy-five million dollars upfront, up to $1.53 billion, global rights. Also preclinical. So $185 million of Western cash went into two Chinese molecules in one week, neither of which has a human dose behind it. Preclinical assets used to be the part of a Chinese pipeline a buyer got thrown in with the package. They are now the thing being bought, which means the diligence sits on the platform and the chemistry rather than on a response rate, and the buyer is paying for a thesis it cannot yet check. $HCM $GSK $2096 HK $RHHBY $XBI $IBB ✨ Made with AI

Boan Biotech has China clearance to start human trials of BA2201, a TL1A x IL-23p19 bispecific for inflammatory bowel disease. Announced 6 Nov; the IND was accepted in August. The plan: a randomized, double-blind, placebo-controlled Phase I/IIa, dosed under the skin, in healthy volunteers and patients with moderate-to-severe ulcerative colitis. Boan says monkey half-life suggests dosing every three months may be possible. No human data yet. Same target pair as Caldera's CLD-423 (from Qyuns, in Phase 1 in Australia and China) and InnoCare's ICP-B161, first dosed in China on 30 Oct. That makes at least three China-origin IL-23 x TL1A bispecifics in or entering the clinic. $6955 HK $2509 HK $9969 HK $XBI $IBB ✨ Made with AI

AusperBio's hepatitis B antisense drug AHB-137: in the NA-treated cohort, functional cure held in 16 of 17 evaluable patients 12 months after stopping all HBV treatment, and in 9 of 9 at 15 months. Company-reported AASLD 2026 data, oral presentations on 8 Nov. NA-treated cohort of 34 (21 complete responders). In treatment-naive patients, 6 of 7 complete responders held through 15 months; one rebounded. The evaluable count shrinks with follow-up, and safety is described only in words (no severe flares or new signals), with no figures. Why it matters: GSK's bepirovirsen is the same drug class, an antisense oligonucleotide. AusperBio, a venture-backed China/US company, already has a Phase 3 running in China. $GSK $XBI $IBB ✨ Made with AI

Adding Brii Bio's therapeutic vaccine and an siRNA to interferon more than doubled hepatitis B surface antigen loss: 25.5% (25 of 98) vs 10.2% (5 of 49) on interferon alone at end of treatment, p=0.0301. Late-breaking AASLD data released 5 Nov, Phase 2 ENHANCE, company-reported, virologically suppressed adults. 24 weeks off treatment: 17.3% vs 6.1%. In patients starting above 1,000 IU/mL HBsAg: 8 of 33 vs 0 of 18. The siRNA is elebsiran, licensed from Vir for Greater China in 2020. Why it matters: the control is interferon alone, the existing finite-course route to functional cure that any new regimen would have to beat in China. $2137 HK $VIR $XBI $IBB ✨ Made with AI

Antengene has dosed the first patient with ATG-201, a CD19xCD3 T-cell engager, in B cell-driven autoimmune disease. Announced 6 Nov. Phase 1 ATTRACT: dose escalation, then expansion, in China and Australia. Company release; no n, registry number or dosing date given. Lead investigator: Zhanguo Li, Peking University People's Hospital. UCB holds worldwide rights to develop, make and sell it, so a China-run Phase 1 is now generating the first human data for a Belgian pharma's asset. The read-through: off-the-shelf CD19 engagers are the alternative to autologous CAR-T in autoimmunity, where Novartis has paused its autoimmune CAR-T trials since late August after three patient deaths. $6996 HK $NVS $XBI $IBB ✨ Made with AI

Lynk Pharmaceuticals has filed its oral JAK1 inhibitor zemprocitinib in China for rheumatoid arthritis after biologics fail. The CDE accepted the NDA on 4 Nov. The evidence: a China Phase 3, double-blind, placebo-controlled, n=430, in patients with inadequate response or intolerance to biologic DMARDs. Company-reported: primary and key secondary endpoints met, p<0.0001, with no arm-level numbers released. On 76 weeks of continuous treatment: ACR20 91.8%, ACR50 80.5%, ACR70 62.4%, denominator not given. Lynk is private, with a Hong Kong listing application pending; Simcere holds exclusive China commercial rights in rheumatology. An atopic dermatitis NDA has been under review since April. Why it matters: biologic failures are the population AbbVie tested Rinvoq, also a selective JAK1, in SELECT-BEYOND. China is getting a domestic option for the same patients. $2096 HK $ABBV $XBI $IBB ✨ Made with AI

Hengrui has China clearance to test its HER2 ADC trastuzumab rezetecan (SHR-A1811) with a PD-1/PD-L1 inhibitor around surgery, in resectable non-squamous lung cancer with HER2 abnormalities. Source: filing no. 2026-170, signed 4 Nov, acceptance CXSL2600810. The same filing lists a notice for Hengrui's PD-L1 adebrelimab, CXSL2600811. No phase, n or design given. The ADC's lung label today is HER2-mutant disease after at least one systemic therapy. So in lung, the ADC is moving from later-line toward curative-intent treatment. An investigator-led neoadjuvant Phase 2 with adebrelimab (NCT07281209, n=40) is already registered. Cumulative R&D on the ADC: about RMB 2.36B, unaudited. The filing puts 2025 global sales of four HER2 ADCs, Enhertu among them, at about $8.16B. $600276 SH $1276 HK $AZN $DSNKY $XBI $IBB ✨ Made with AI

BeOne's US Brukinsa sales last quarter, about $1.0B, were roughly nine times Zai Lab's whole revenue. Both reported on 5 Nov. BeOne: Q3 revenue $1.89B, up 34%; GAAP net income $385M; 2026 revenue guidance raised to $6.9-7.0B from $6.6-6.8B. Global Brukinsa $1.4B, Tevimbra $243M. Zai: Q3 revenue $108.2M vs $116.1M; net loss $46.1M, with $13.6M of inventory write-downs. It sells licensed drugs mainly in China and says that business returns to growth in 2027. Zai's bet is outside China: its DLL3 ADC zoci, with a global Phase 3 in first-line small cell lung cancer due to start by year-end, and a possible first US BLA in 2027. It raised $250M in October at $22.04 per ADS. What I'm watching: zoci's Phase 3 start, then whether the interim from DLLEVATE, its Phase 3 in second-line-plus SCLC, lands in 2H 2027. $ONC $ZLAB $XBI $IBB ✨ Made with AI

Ascentage's first clinical look at its oral EED inhibitor APG-5918: ORR 44.7% in r/r T-cell lymphoma (21 of 47, 2 CRs), and 65.0% at the 600 mg daily dose it chose for Phase 2 (n=20, CR 10%). Source: the 4 Nov ASH preview; oral #513 on 13 Dec. First-in-human Phase 1, US and China, company-reported, enrollment cutoff 30 July. No DLTs; grade 3+ treatment-related events mainly hematologic. DOR and PFS not reached. The yardstick is Daiichi Sankyo's valemetostat, an EZH1/2 inhibitor on the same PRC2 complex. Lancet Oncology, Phase 2, r/r PTCL: ORR 44% (52 of 119), CR 14%, median DOR 11.9 months. Different populations, and 20 patients at one dose is thin. What I'm watching: durability at the December oral. $AAPG $6855 HK $DSNKY $XBI $IBB ✨ Made with AI

Aosaikang has filed ASKB589 (omectatug), its CLDN18.2 antibody, in China for first-line gastric cancer, on a Phase 3 it says hit PFS. Acceptance no. CXSS2600184, announced 4 Nov. Chinese trade press calls it the first domestic CLDN18.2 antibody to file. The Phase 3 added it to BeOne's tislelizumab plus CAPOX, against placebo on the same backbone. Company statement: statistically significant IRC-assessed PFS, consistent benefit on OS. No HR, months or n disclosed. The ESMO 2025 Phase Ib/II (62 enrolled, PD-1 plus CAPOX): 47 stage IV moderate-to-high expressers had cORR 76.1%, median PFS 12.45 months. Astellas' Vyloy, the first approved CLDN18.2 antibody, carries its China label with chemo. A PD-1 combination filing would test a different backbone in overlapping patients. China also has CARsgen's later-line CAR-T approved and CLDN18.2 ADCs under review. What I'm watching: the Phase 3 hazard ratio. $002755 SZ $ALPMY $ONC $2171 HK $XBI $IBB ✨ Made with AI

Legend Biotech's in-vivo CAR-T LB2501: all 8 patients at dose level 2 have reached complete response, per its 4 Nov ASH preview. June's set (1 April cutoff) was 5 of 6. Company-reported, China Phase 1 (NCT07002112), r/r B-cell lymphoma, responses lasting up to five months. The release gives no cutoff date and no LB2501 safety numbers; poster #4169 comes on 13 Dec. June's set: no lymphodepleting chemo, DL2 CRS in 6 of 6, all grade 1-2, no ICANS, and 0 of 6 responses at the lower dose. Eight patients and five months is still thin. The number to look for in December: how many of the 8 are still in CR. $LEGN $XBI $IBB ✨ Made with AI

JUST IN: Hinova licensed its oral VAV1 molecular glue HP570 (EVO570 at Evommune) outside Greater China: $30M upfront, up to about $1.07B in milestones, royalties from mid-single to low-double digits. Chinese headlines say $30M. Evommune's 8-K of 4 Nov splits it: $20M cash plus pre-funded warrants on 1,262,413 shares, valued at $10M on a 45-day average implying about $7.92. Evommune's last price that day was $6.04, the price at which it agreed to sell $30M of stock to investors. Hinova can instead ask for cash at 90% of that average, about $9M. The benchmark: Novartis paid Monte Rosa $150M upfront and up to $2.1B in milestones in October 2024 for MRT-6160, then in Phase 1, now in Phase 2. Different structures: Novartis took worldwide rights with a US profit share. Hinova got a fifth of that upfront with no reported human data: IND accepted in the US, submitted in China, first Phase 2 in rheumatoid arthritis planned for 2027. $688302 SH $EVMN $GLUE $NVS $XBI $IBB ✨ Made with AI