Rough­ly three years ago, As­traZeneca teamed up with Lon­don’s Benev­o­len­tAI to bring new drugs in­to its port­fo­lio us­ing the biotech’s AI and ma­chine learn­ing ca­pa­bil­i­ties. Now that the orig­i­nal deal has borne fruit, the two com­pa­nies are re-up­ping their col­lab­o­ra­tion.

As­traZeneca and Benev­o­len­tAI will work to­geth­er in two more dis­ease ar­eas: sys­temic lu­pus and heart fail­ure, the com­pa­nies an­nounced Thurs­day. It’s a three-year part­ner­ship that comes af­ter two can­di­dates dis­cov­ered by Benev­o­len­tAI for chron­ic kid­ney dis­ease and id­io­path­ic pul­monary fi­bro­sis were nom­i­nat­ed to As­traZeneca’s port­fo­lio last year.

No fi­nan­cials were dis­closed, but Benev­o­len­tAI CSO Anne Phe­lan told End­points News there would be an up­front pay­ment, an eq­ui­ty in­vest­ment and mile­stones.

The orig­i­nal part­ner­ship proved a sig­nif­i­cant mile­stone for As­traZeneca, as the CKD mol­e­cule nom­i­nat­ed last Jan­u­ary was the com­pa­ny’s first gen­er­at­ed by AI. At the time, As­traZeneca iden­ti­fied and val­i­dat­ed a new way to at­tack CKD thanks to the biotech’s plat­form AI tech, and be­gan de­vel­op­ing com­pounds cen­tered on this tar­get.

The IPF com­pound fol­lowed up just last month, and now marks the “per­fect time” to ex­tend the work­ing re­la­tion­ship, Phe­lan said.

Lu­pus and heart fail­ure “fall with­in [As­traZeneca’s] pri­or­i­ty ther­a­peu­tic ar­eas,” Phe­lan said. “They’re com­plex dis­eases, dif­fi­cult spaces to work in. It’s a tricky prob­lem that lends it­self well to AI and our ma­chine learn­ing ca­pa­bil­i­ties.”

At the heart of Benev­o­len­tAI’s tech­nol­o­gy is what Phe­lan de­scribes as a “knowl­edge graph,” which es­sen­tial­ly looks like a “hair­ball” net­work of thou­sands of da­ta points, or nodes, in­ter­con­nect­ed by mil­lions of lines or re­la­tion­ships. Such re­la­tion­ships are first iden­ti­fied by the ma­chine learn­ing al­go­rithms and can con­nect many dif­fer­ent and un­re­lat­ed dis­eases.

Anne Phe­lan

Benev­o­len­tAI is able to dif­fer­en­ti­ate it­self from the buzzy and grow­ing field of AI drug de­vel­op­ment be­cause it is dri­ven by a hy­poth­e­sis-first ap­proach rather than fo­cus­ing on phe­no­typ­ic screen­ing, Phe­lan added. Where­as com­peti­tors’ AI might be mak­ing new mol­e­cules and try­ing to re­verse-en­gi­neer the process, Benev­o­len­tAI starts from scratch.

“It’s not a right and wrong sit­u­a­tion … it’s a re­al­ly mul­ti­di­men­sion­al pic­ture of hu­man bi­ol­o­gy across the com­mon­al­i­ties of hu­man dis­ease,” Phe­lan said. “For ex­am­ple, de­pend­ing on the cell type that’s af­fect­ed, the graph could end up con­nect­ing to Alzheimer’s or di­a­betes, where un­der­ly­ing phys­i­ol­o­gy could be re­al­ly quite sim­i­lar.”

For As­traZeneca, lu­pus re­cent­ly turned in­to a po­ten­tial mon­ey­mak­er, as the Big Phar­ma whisked a new drug past the FDA fin­ish line last Au­gust in Saph­ne­lo. It’s the first new lu­pus ap­proval in at least the last 10 years, As­traZeneca said at the time, and re­duces over­all dis­ease ac­tiv­i­ty across or­gan sys­tems and oral cor­ti­cos­teroid use com­pared to place­bo.

Heart fail­ure, mean­while, is an area where As­traZeneca is look­ing to stake out a block­buster with Farx­i­ga. Back in 2020, the drug was ap­proved to re­duce the risk of car­dio­vas­cu­lar death or hos­pi­tal­iza­tion in heart fail­ure pa­tients with a re­duced ejec­tion frac­tion. It al­so picked up an ap­proval in CKD last April.

Belgian biotech Confo Therapeutics has landed $183 million, plus potential royalties, in a drug-discovery deal with Daiichi Sankyo.

Early Thursday, Confo Therapeutics put out word of the deal that will be focused on small molecule antagonists to go after an undisclosed target that the company says is associated with CNS diseases.

Confo CEO Cedric Ververken told Endpoints News that Daiichi originally reached out to learn about the biotech’s technology. He added that Confo, founded in 2015, will use its platform to drug a GPCR target that Daiichi has struggled with internally.

Otonomy may be shutting down, but the lessons learned there will live on at another biotech working on new treatments for hearing loss.

San Francisco-based Spiral Therapeutics has bought certain assets related to three of Otonomy’s programs, ranging from data, patent rights, and know-how to inventory. That includes data around Otonomy’s twice-failed lead program, OTO-104 (Otividex), a sustained-exposure formulation of dexamethasone.

Paul McKenzie took over as CEO of Australian pharma giant CSL this month, following in the footsteps of long-time CSL vet Paul Perreault.

With an eye on mRNA, and quickly commercializing its new, $3.5 million-per-shot gene therapy for hemophilia B, McKenzie and chief medical officer Bill Mezzanotte answered some questions from Endpoints News this afternoon about where McKenzie is going to take the company and what advances may be coming to market from CSL’s pipeline. Below is a lightly edited transcript.

Boehringer Ingelheim reported human pharma sales of €18.5 billion on Wednesday, led by type 2 diabetes and heart failure drug Jardiance and pulmonary fibrosis med Ofev. Jardiance sales reached €5.8 billion, growing 39% year over year, while Ofev took in €3.2 billion, notching its own 20.6% annual jump.

However, Boehringer is also looking ahead with its pipeline, estimating “In the next seven years the company expects about 20 regulatory approvals in human pharma.”