Omid Farokhzad (file photo)

Seer rais­es $55M as pro­teomics ap­proach­es prime time in dis­ease de­tec­tion, drug de­vel­op­ment

Omid Farokhzad was work­ing in his Har­vard nan­otech lab in 2017 when he found some­thing he knew would be­come his next com­pa­ny. Two years and, as of to­day, over $100 mil­lion in fund­ing for lat­er, he still hasn’t re­vealed what it is.

“We’ve seen them re­port­ed, but we don’t re­al­ly know what their tech­nol­o­gy is or what it does.” Stephen Williams, the CMO of So­ma­Log­ic, the most high-pro­file po­ten­tial com­peti­tor to Farokhzad’s com­pa­ny, Seer, told End­points News.

The tech­nol­o­gy was big, though, or at least Farokhzad thought so. Big enough for him to leave much of his work in Boston, where he launched two com­pa­nies, opened his own lab and spent 20 years across Har­vard and MIT, to spend most of his time build­ing Seer in Red­wood City, out­side San Fran­cis­co. What he said was this: They were try­ing to best-of-both-worlds a sci­en­tif­ic idea. They want­ed to take a study of hu­man pro­teins they said was of­ten con­duct­ed with ei­ther slip­shod speed or painstak­ing pre­ci­sion and make it both fast and metic­u­lous.

Now, ex­act­ly a year af­ter its pub­lic launch and just as So­ma­Log­ic eyes its biggest com­mer­cial ex­pan­sion, Seer is an­nounc­ing $55 mil­lion in Se­ries D fund­ing and its own piv­ot from a re­search-fo­cused com­pa­ny to one try­ing to bring the pro­teome in­to the med­ical main­stream.

“There was the mo­ment, where I said I need to do this my­self,” Farokhzad told End­points, “and if I don’t I’m go­ing to re­gret it for the rest of my life.” 

Seer, So­ma­Log­ic, and a hand­ful of oth­er com­pa­nies work on what’s be­come known as the pro­teome. For decades, it had been tossed around as a bi­o­log­i­cal what-if, as se­duc­tive as it was un­at­tain­able. Per­haps in­stead of look­ing at DNA and RNA to un­der­stand the hu­man body and its dis­or­ders, we could look at the pro­teins that ge­net­ic code makes, the ones that in most cas­es ul­ti­mate­ly do the dam­age.

The hy­po­thet­i­cals were huge. In­stead of pre­dict­ing, based on ge­net­ics, if a 50-year-old had a pre­dis­po­si­tion to heart at­tacks or di­a­betes, a doc­tor could check if there had been ac­tu­al changes in the blood that built to those ail­ments. Phar­ma com­pa­nies could bet­ter un­der­stand how their drugs worked, and re­fine ac­cord­ing­ly.

One ob­sta­cle was math­e­mat­i­cal. With pro­teins, there are sim­ply far more pos­si­bil­i­ties. DNA has four let­ters and bonds in very spe­cif­ic ways and struc­tures: A-T, G-C, the dou­ble he­lix. Pro­teins are formed from up to 20 amino acids and can bond in myr­i­ad ways, and then change af­ter their cre­ation in­to a va­ri­ety of states de­pend­ing on what bonds to them.

“You ac­tu­al­ly need a lot of com­put­ing pow­er to cal­cu­late the the­o­ret­i­cal com­plex­i­ty you can cre­ate with pro­teins,” Farokhzad said. “You’ll nev­er get there.”

The oth­er prob­lem was chem­i­cal. DNA and RNA can be repli­cat­ed and am­pli­fied in a lab. Pro­teins can­not. In­stead, re­searchers have re­lied on bi­o­log­i­cal fish­ing ex­pe­di­tions. First, they used an­ti­bod­ies, which bind to pro­teins and make them eas­i­er to spot. But each an­ti­body on­ly binds to one pro­tein. Us­ing them to iden­ti­fy every hu­man pro­tein would be, as the New York Times’ Michael Be­har put it last year, “like try­ing to cat­a­log every fish in the ocean with a net that cap­tured on­ly a sin­gle species at a time.”

So­ma­Log­ic built it­self on the dis­cov­ery that nu­cle­ic acids called ap­tamers could be used to iden­ti­fy far more pro­teins at once. They built a data­base of 5,000 hu­man pro­teins, and then use ma­chine learn­ing to see how changes in those pro­teins’ struc­tures and quan­ti­ty cor­re­late with cer­tain health fac­tors. Af­ter decades of de­vel­op­ment and years of clin­i­cal tests, they launched their So­maSig­nal for med­ical prac­tices in Sep­tem­ber and bur­nished it with a De­cem­ber study in Na­ture that scanned 17,000 par­tic­i­pants for those 5,000 pro­teins and pre­dict­ed for di­a­betes and heart at­tack, among oth­er health is­sues.

Those who test­ed as high-risk could then be re­ferred to pre­ven­ta­tive mea­sures they may not oth­er­wise have re­ceived.

Stephen Williams came to So­ma­Log­ic in 2009 af­ter near­ly 20 years in ex­per­i­men­tal and clin­i­cal de­vel­op­ment at Pfiz­er

Click on the im­age to see the full-sized ver­sion

“We tried to de­vel­op 13 mod­els for 13 health is­sues, and we suc­ceed­ed with 11,” Williams said, adding they failed on­ly on weight pre­dic­tions and cur­rent di­et. “We re­al­ly an­swered the ques­tion: Can pro­teins be­come a sole in­for­ma­tion source?”

Farokhzad’s cri­tique is that this is a ‘bi­ased’ ap­proach: So­ma­Log­ic can on­ly look for the pro­teins it has ap­tamers for and it al­so can’t look deeply in­to spe­cif­ic changes. He said his Seer will be ‘un­bi­ased,’ rapid­ly iden­ti­fy­ing every pro­tein – at least 20,000 – and all the changes that can oc­cur with­in them and then use their own ma­chine learn­ing to find pat­terns. “Fast and broad,” he likes to call it.

“To re­al­ly un­der­stand pro­teomics, you have to do it in an un­bi­ased ap­proach,” he said.  “New bi­ol­o­gy is al­ways dri­ven by what we don’t know yet.”

Out in Red­wood City, Farokhzad has set up a Seer head­quar­ters with 40 em­ploy­ees and plans to ex­pand to 80 with the new fund­ing. He said they’ve spent the last year with their heads down try­ing to per­fect the sci­ence. Afraid biobanks might con­t­a­m­i­nate sam­ples with poor pro­ce­dures, they’ve set up 70 clin­i­cal sites na­tion­wide to build their data­base.

Oth­er folks have used an un­bi­ased ap­proach – help­ing build data­bas­es like the Hu­man Pro­teome Pro­ject, which now counts 19,823 pro­teins – but comb­ing through all those amino acids takes for­ev­er. Farokhzad has promised to make it fast enough to work with drug com­pa­nies (they al­ready have part­ners) and un­veil scans like those of­fered by So­ma­Log­ic.

If the tech holds up, it would have big im­pli­ca­tions for ear­ly de­tec­tion, un­veil­ing pre­vi­ous­ly hid­den changes in the body that presage dis­ease. Farokhzard claims it could help clin­i­cal tri­als, nam­ing Alzheimer’s and the hy­poth­e­sis that some of the drugs that failed might work if they treat a pa­tient ear­ly – at a stage we can’t yet de­tect. Like many in pro­teomics, he talks about it like a new fron­tier.

“We’re ba­si­cal­ly open­ing the flood­gate of ac­cess to pro­teom­ic in­for­ma­tion,” he said.

In So­ma­Log­ic’s view, though, that grand vi­sion isn’t nec­es­sary. All ap­proach­es are bi­ased, Williams said. The ques­tion is whether you can make pre­dic­tions that can be put to med­ical use.

“The ques­tion is whether you can mea­sure some­thing that is use­ful or ac­tion­able, not whether there’s some in­her­ent truth,” he said. “For our busi­ness, we don’t need to un­der­stand what the bi­ol­o­gy is. We let the ma­chine learn­ing pick out the best com­bi­na­tion of mea­sure­ments.”

Still, Williams said that when their long­time part­ner No­var­tis – which us­es So­ma­Log­ic’s tech to un­der­stand the mech­a­nisms of their drugs, among oth­er things – needs a high­er res­o­lu­tion look at what So­ma­Log­ic tech­nol­o­gy turns up, they di­rect them to a dif­fer­ent plat­form. Seer is de­vel­op­ing a com­mer­cial scan like So­ma­Log­ic’s and, like So­ma­Log­ic, even­tu­al­ly hopes to put out a (some­what con­tro­ver­sial) chip any­one can use, but their eas­i­est niche may be in drug de­vel­op­ment.

The in­dus­try will know soon enough. So­maSig­nal is on­ly avail­able in Col­orado for now, but a big­ger roll­out is pend­ing and they just hired a new com­mer­cial of­fi­cer. Af­ter two years in rel­a­tive ob­scu­ri­ty, Farokhzad said Seer will soon be open­ing up, and a prod­uct launch is sched­uled for 2021.

“Our tech­nol­o­gy risk is now large­ly put to rest and what lays ahead of us is ba­si­cal­ly prod­uct de­vel­op­ment and com­mer­cial­iza­tion,” he said.  “Now we’ve got to ex­e­cute.”

Stephen Hahn, FDA commissioner (AP Images)

As FDA sets the stage for the first Covid-19 vac­cine EUAs, some big play­ers are ask­ing for a tweak of the guide­lines

Setting the stage for an extraordinary one-day meeting of the Vaccines and Related Biological Products Advisory Committee this Thursday, the FDA has cleared 2 experts of financial conflicts to help beef up the committee. And regulators went on to specify the safety, efficacy and CMC input they’re looking for on EUAs, before they move on to the full BLA approval process.

All of this has already been spelled out to the developers. But the devil is in the details, and it’s clear from the first round of posted responses that some of the top players — including J&J and Pfizer — would like some adjustments and added feedback. And on Thursday, the experts can offer their own thoughts on shaping the first OKs.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 92,200+ biopharma pros reading Endpoints daily — and it's free.

A new chap­ter in the de­cen­tral­ized clin­i­cal tri­al ap­proach

Despite the promised decentralized trial revolution, we haven’t yet moved the needle in a significant way, although we are seeing far bolder commitments to this as we continue to experience the pandemic restrictions for some time to come. The vision of grandeur is one thing, but operationalizing and execution are another and recognising that change, particularly mid-flight on studies, is worthy of thorough evaluation and consideration in order to achieve success. Here we will discuss one of the critical building blocks of a Decentralized and Remote Trial strategy: TeleConsent; more than paper under glass, it is a paradigm change and key digital enabler.

Michel Vounatsos, Biogen CEO (via YouTube)

UP­DAT­ED: Bio­gen spot­lights a pair of painful pipeline set­backs as ad­u­canum­ab show­down looms at the FDA

Biogen has flagged a pair of setbacks in the pipeline, spotlighting the final failure for a one-time top MS prospect while scrapping a gene therapy for SMA after the IND was put on hold due to toxicity.

Both failures will raise the stakes even higher on aducanumab, the Alzheimer’s drug that Biogen is betting the ranch on, determined to pursue an FDA OK despite significant skepticism they can make it with mixed results and a reliance on post hoc data mining. And the failures are being reported as Biogen was forced to cut its profit forecast for 2020 as a generic rival started to erode their big franchise drug.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 92,200+ biopharma pros reading Endpoints daily — and it's free.

UP­DAT­ED: CRISPR Ther­a­peu­tics gets a snap­shot of off-the-shelf CAR-T suc­cess in B-cell ma­lig­nan­cies — marred by the death of a pa­tient

Just days after scientific founder Emmanuelle Charpentier shared the Nobel prize for her work on CRISPR/Cas9, CRISPR Therapeutics $CRSP is showing off a snapshot of success in their early-stage study for an off-the-shelf CAR-T approach to CD19+ B cell malignancies — a snapshot marred by the death of a patient who had been given a high dose of the treatment.

Using their gene editing tech, researchers for CRISPR engineered cells from healthy donors into an attack vehicle aimed at cancer, something that has been achieved with great success using patients’ own cells — the autologous approach. But autologous CAR-T is hampered by the more complex vein-to-vein requirement that delays treatment, and now CRISPR Therapeutics along with other players like Allogene are determined to replace the pioneers with CAR-T 2.0.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 92,200+ biopharma pros reading Endpoints daily — and it's free.

RBC's Bri­an Abra­hams holds a mock ad­comm on Bio­gen's iffy ad­u­canum­ab da­ta — and most of these ex­perts don't see a path to an ap­proval

As catalysts go, few loom larger than the aducanumab adcomm slated for Nov. 6.

With its big franchise under assault, Biogen is betting the ranch that its mixed late-stage Alzheimer’s data can squeak past the experts and regulators and get onto the market. And the topic — after a decade of Alzheimer’s R&D disasters in what still represents the El Dorado of drug markets — remains in the center ring of discussions around late-stage pipeline prospects.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 92,200+ biopharma pros reading Endpoints daily — and it's free.

David Hung (file photo)

Mas­ter deal­mak­er David Hung re­tools a SPAC sedan in­to a fi­nanc­ing mus­cle ve­hi­cle that leaves his can­cer start­up with $850M and a place on Wall Street

It’s only right that one of the industry’s top dealmakers just completed one of the biggest SPAC-related deals in the pipeline.

David Hung, of Medivation fame, has completed a back flip into the market, merging with EcoR1 Capital’s SPAC Panacea and landing neatly on Wall Street with an $NUVB stock ticker after filling out the blank check in his name. In addition to the $144 million held in the SPAC — provided none of the investors opt out — Hung is getting ahold of $500 million more being chipped in by a slate of institutional investors who feel that Hung could have the keys to another Medivation-style success.

Endpoints News

Keep reading Endpoints with a free subscription

Unlock this story instantly and join 92,200+ biopharma pros reading Endpoints daily — and it's free.

Years af­ter a ma­jor tri­al set­back, No­var­tis switch­es gears with SMA drug. This time they're try­ing it for Hunt­ing­ton's

Four years after a Phase I/II setback in spinal muscular atrophy (SMA), Novartis is hoping its drug branaplam will find more success in a new neurological indication: Huntington’s disease.

The decision was announced a year after the head of research, Jay Bradner, said he did not see a “big opportunity” in SMA, according to Reuters. Novartis says it has preclinical data showing that branaplam reduces levels of mutant huntingtin protein, and SMA data showing patients on the drug had reductions in huntingtin mRNA. The FDA gave branaplam their orphan drug designation, and Novartis plans to move forth with a Phase IIb trial next year.

Glax­o­SmithK­line's vac­cines group aims for a first as it kicks off PhI­II RSV stud­ies

One of GlaxoSmithKline’s big projects at its global vaccine R&D center in Rockville, MD is set to enter Phase III after passing early-stage tests with flying colors.

Eyeing the wide-open respiratory syncytial virus (RSV) space, GSK is pushing two different vaccine candidates: GSK3888550A is designed to confer protection to infants via maternal immunization, while GSK3844766A is meant for the elderly.

Pur­due Phar­ma signs guilty plea, preps $8B+ set­tle­ment on Oxy con­tro­ver­sy — re­port; Flag­ship brings in a comms chief

Purdue Pharma may soon be signing off on a guilty plea and an $8 billion-plus settlement to wrap up its controversial role distributing OxyContin.

The AP has the breaking story this morning.

Purdue filed for bankruptcy last year, along with Insys and followed by Mallinckrodt, as it navigated its way through a blizzard of litigation surrounding Oxy, which triggered an epidemic of abuse around the country.