It was the cel­lu­lar equiv­a­lent of open­ing your car door and find­ing an ac­tive, roar­ing en­gine in the dri­ver seat.

Paul Mis­chel

Sci­en­tists learned strands of DNA could oc­ca­sion­al­ly ap­pear out­side of its tra­di­tion­al home in the nu­cle­us in the 1970s, when they ap­peared as lit­tle, in­nocu­ous cir­cles on mi­cro­scopes; in­ex­plic­a­ble but ap­par­ent­ly in­nate. But not un­til UC San Diego’s Paul Mis­chel pub­lished his first study in Sci­ence in 2014 did re­searchers re­al­ize these cir­cles were not on­ly ac­tive but po­ten­tial­ly over­ac­tive and dri­ving some can­cer tu­mors’ su­per­hu­man growth.

That in­sight and the en­su­ing five years of re­search will now get $46 mil­lion cash and com­pa­ny in­fra­struc­ture to ramp in­to tar­get­ed ther­a­pies as Bound­less Bio emerges from stealth mode with back­ing from ARCH Ven­ture Part­ners and City Hill. Ques­tions abound, from what pre­cise­ly a drug would look like to what even gives rise to these wild DNA, but CEO Zachary Horn­by isn’t bit­ing his tongue on the po­ten­tial.

“We’re think­ing about this as a fourth rev­o­lu­tion in can­cer ther­a­pies,” Horn­by, who was most re­cent­ly COO of Igny­ta, told End­points News. The first three rev­o­lu­tions, by Horn­by’s count, are chemother­a­py in the 1940s, the first tar­get­ed ther­a­pies at the end of the 20th cen­tu­ry, and the re­cent rise of im­munother­a­py.

The road to such a rev­o­lu­tion would be long, but the em­bat­tled on­col­o­gy field may be in need of new di­rec­tion. A study re­leased in April found 97% of can­cer drugs test­ed in clin­i­cal tri­als failed to make it to mar­ket, and this month re­searchers found sys­temic tar­get­ing prob­lems plagued two decades of can­cer re­search.

The con­nec­tion be­tween this loose DNA, of­fi­cial­ly called ex­tra­chro­mo­so­mal DNA or ecD­NA, and can­cer cen­ters around a baf­fling and dead­ly fact about some tu­mors: While nor­mal cells in a high­er-or­der species like hu­mans don’t evolve with­in a gen­er­a­tion, some can­cer cells can evolve rapid­ly, en­sur­ing their sur­vival against at­tempt­ed treat­ments. Why? How? Mis­chel’s map­ping of can­cer genome points to ecD­NA.

Freed from a cell’s chro­mo­somes, the DNA can repli­cate rapid­ly. That doesn’t hurt if they code for noth­ing or some­thing be­nign, but if they code for some­thing that gives the cell an ad­van­tage, such as EGFR (a growth fac­tor), the cells will grow rapid­ly as in any clas­si­cal nat­ur­al se­lec­tion mod­el. This, Horn­by said, ap­pears in over 25% of can­cers, in­clud­ing no­to­ri­ous­ly hard to treat MET can­cers.

EGFR in­hibitors al­ready ex­ist to com­bat can­cer cells that have al­ready evolved (or been “am­pli­fied”), but Bound­less Bio plans to use Mis­chel’s in­sights to de­stroy ecD­NA in its ear­ly stages. Rather than at­tack­ing tu­mors af­ter the cells have al­ready am­pli­fied, the com­pa­ny would jam the process that gives rise to the evo­lu­tion in the first place.

“It opens a whole new av­enue of can­cer tar­gets, in­clud­ing al­low­ing us to pur­sue pa­tient pop­u­la­tions that to this point have been un­drug­gable,” Horn­by said, point­ing to MET and Myc. “That’s just a re­al­ly dif­fer­ent ap­proach than your typ­i­cal tar­get­ed ther­a­pies.”

But how they would do this is still cloudy.

Horn­by said the most promis­ing method was jam­ming the “en­zy­mat­ic ma­chin­ery” — the mol­e­c­u­lar tools and parts that al­low DNA to repli­cate and code pro­teins — as their re­search has shown the ma­chin­ery is slight­ly dif­fer­ent in ecD­NA than typ­i­cal DNA. An­oth­er method they’re ex­plor­ing is to in­hib­it the meta­bol­ic path­ways ecD­NA can use to ful­fill the de­mands caused by its high repli­ca­tion rate; in oth­er words, grow­ing DNA that are hun­gry and de­priv­ing them of food could neu­tral­ize them.

Among the most no­table things about Bound­less’ po­ten­tial ther­a­pies is that they may be ap­proved for tu­mor type, rather than can­cer type, i.e. like the new drug from Horn­by’s old com­pa­ny Igny­ta, it could treat a wide range of can­cers if the pa­tient showed ecD­NA was am­pli­fy­ing in the tu­mor.

The com­pa­ny will al­so in­vest in re­search to dis­cov­er the un­der­ly­ing mech­a­nism giv­ing rise to ecD­NA.

89bio said its drug was better than placebo at lessening fibrosis without worsening nonalcoholic steatohepatitis, or NASH, in two of three dose groups.

The San Francisco biotech said it thinks the Phase IIb data pave the way for a potential Phase III, following in the footsteps of another biotech in its drug class, Akero Therapeutics. To fund a late-stage study, CEO Rohan Palekar told Endpoints News 89bio “would need to raise additional capital,” with the company having about $188 million at the end of last year.