π Health and biotech
How to overcome the growing antibiotic resistance problem

The post-antibiotic era: a health disaster is predicted

Agnès Vernet, Science journalist
On June 16th, 2022 |
5 min reading time
Hannu Myllykallio
Hannu Myllykallio
CNRS research director at LOB* at École Polytechnique (IP Paris)
Michael Mourez
Michael Mourez
Director of Innovation at DEINOVE
Key takeaways
  • In 2015, the O’Neill report showed that antibiotic resistance could kill more people than cancer by 2050.
  • The traditional approach of analysing the contents of an extract that kills the bacteria in the laboratory no longer works.
  • Some laboratories are trying to use phages, specific viruses, bacteria to create new therapeutic approaches.
  • As each infection is biologically unique, the meeting of a strain, host and context becomes more difficult.

New sci­en­tif­ic approach­es, new clin­i­cal eval­u­a­tion frame­works, new eco­nom­ic mod­els… What if, to solve the antibi­ot­ic resis­tance cri­sis, we had to rein­vent everything?

Faced with the devel­op­ment of resis­tance to antimi­cro­bial treat­ments the World Health Organ­i­sa­tion is warn­ing that we are on the verge of enter­ing a post-antibi­ot­ic era and that a health cat­a­stro­phe is on the hori­zon. Han­nu Myl­lykallio, pro­fes­sor at the Optics and Bio­sciences lab­o­ra­to­ry (LOB) at École Poly­tech­nique, reminds us that “the O’Neill report writ­ten by econ­o­mist Jim O’Neill on behalf of the British gov­ern­ment1 showed that antibi­ot­ic resis­tance could kill more peo­ple than can­cer by 2050.”

So, has research giv­en up hope of devel­op­ing new solu­tions to bac­te­r­i­al dis­eases? No. “A great many aca­d­e­m­ic lab­o­ra­to­ries are work­ing on the sub­ject,” con­firms Han­nu Myl­lykallio. “But to suc­ceed, many obsta­cles must be over­come.” For Michael Mourez, direc­tor of inno­va­tion at Mont­pel­li­er-based biotech Deinove, which invents new antibi­otics by com­bin­ing syn­thet­ic biol­o­gy and analy­ses of lit­tle-known bac­te­ria, “the devel­op­ment of antibi­otics faces three types of chal­lenge: sci­en­tif­ic, phar­ma­co­log­i­cal and economic.”

Scientific challenge

While the first is not specif­i­cal­ly an issue of antibi­otics, it is still impor­tant. “The tra­di­tion­al approach, which con­sists of analysing the con­tent of a fer­men­ta­tion extract of a microbe that kills the bac­te­ria in the lab­o­ra­to­ry, has not worked for the past 40 years for gram-neg­a­tive bac­te­ria and 20 years for gram-pos­i­tive bac­te­ria. We need to change the par­a­digm,” Michael Mourez says.

New approach­es have been devel­oped. Many hopes have been pinned on tar­get­ed screen­ing, i.e. the analy­sis of libraries of known mol­e­cules to find new antibi­otics direct­ed against known tar­gets. “But it doesn’t work because phar­ma­col­o­gy of the 20th Cen­tu­ry, which very focused on effi­ca­cy in human cells, does not suf­fi­cient­ly cov­er every bac­te­ria, espe­cial­ly gram-neg­a­tive ones,” explains Michael Mourez. Clear­ly, hav­ing stud­ied bac­te­ria too lit­tle, the phar­ma­ceu­ti­cal indus­try does not under­stand enough about micro­bial biology. 

More­over, genom­ic approach­es would seem to have not kept their promis­es. “We thought it would be enough to iden­ti­fy a new tar­get based on the analy­sis of its gene… but we neglect­ed the trans­la­tion­al aspects,” insists the micro­bi­ol­o­gist. “The way bac­te­ria behave in their host, in a bio­log­i­cal envi­ron­ment that varies from one organ to anoth­er, is some­thing that can­not be repli­cat­ed in a test tube.”

The sci­en­tif­ic com­mu­ni­ty has recog­nised its pow­er­less­ness. In 2007 and 2015, the man­u­fac­tur­ers AstraZeneca2 and GSK3 even pub­lished the results of their fruit­less research. No one can accuse them of hav­ing lacked ambition.

Today, lab­o­ra­to­ries are try­ing new approach­es. “For exam­ple, arti­fi­cial intel­li­gence sys­tems are used to increase the effi­cien­cy of devel­op­ment process­es or to analyse libraries of mol­e­cules,” explains Han­nu Myl­lykallio. These algo­rithms seek to iden­ti­fy the most promis­ing chem­i­cal struc­tures. The mol­e­cules are then syn­the­sised de novo for eval­u­a­tion. In France, the com­pa­ny Iktos is a leader in this approach. The approach is still unprece­dent­ed for antibi­otics, but in 2022 the Amer­i­can com­pa­ny Insil­i­co Med­i­cine entered phase I with an antifi­brot­ic drug designed entire­ly by arti­fi­cial intelligence.

Oth­ers are using phages, virus­es spe­cif­ic to bac­te­ria, to cre­ate new ther­a­peu­tic approach­es. The French com­pa­ny Phere­cy­des is try­ing to do this, with approach­es that com­bine phages and syn­thet­ic biology.

Final­ly, many are focus­ing on antivir­u­lence, which con­sists of dis­arm­ing bac­te­ria and block­ing their harm­ful action rather than try­ing to pre­vent their growth4. Michael Mourez is not get­ting car­ried away, “we don’t yet know if it will work. There is still the ques­tion of the effec­tive­ness of this approach in the human envi­ron­ment.” Han­nu Myl­lykallio agrees, “it is easy to find active mol­e­cules on bac­te­ria in the lab­o­ra­to­ry, but it is more com­pli­cat­ed to show that they are effec­tive and safe in humans. This requires real research efforts. We can no longer rely on the same mech­a­nisms as pre­vi­ous gen­er­a­tions of treatments.”

Clinically tested…

The ques­tion of effi­ca­cy is not the only thing to con­sid­er when devel­op­ing a drug. Michael Mourez explains, “when you have a prod­uct that works in pre­clin­i­cal stud­ies, you then have to find the right dose to admin­is­ter to patients and find the bal­ance between tox­i­c­i­ty and effi­ca­cy. This is not easy with antibi­otics.” It is impor­tant to under­stand that infec­tions are caused by the mul­ti­pli­ca­tion of a large num­ber of micro­bial cells. To erad­i­cate them, it is not sur­pris­ing to use large dos­es of antibiotics.

And find­ing the right dose faces a recur­ring prob­lem in biol­o­gy: imper­fect mod­els. “Mice are not sen­si­tive to the same pathogens as humans,” says Michael Mourez. “So we use ani­mals with defi­cient immune sys­tems to study infec­tions. There are also in vit­ro tests to pre­dict the poten­cy of an antibi­ot­ic, but their results are far too uncor­re­lat­ed with human biol­o­gy.” Here, again, con­ven­tion­al approach­es fail to solve this phar­ma­co­log­i­cal problem.

Final­ly, the clin­i­cal demon­stra­tion of effi­ca­cy is par­tic­u­lar­ly dif­fi­cult for these drugs. “Cur­rent antibi­otics were devel­oped before clin­i­cal rules were imposed. It is almost impos­si­ble today to show that a new prod­uct is more effec­tive than the old ones,” says Michael Mourez. The dif­fi­cul­ties are mount­ing. As each infec­tion is bio­log­i­cal­ly unique, select­ing a uni­form group of patients is vir­tu­al­ly impos­si­ble. As for a com­par­i­son with a place­bo, this is unthink­able from an eth­i­cal point of view, and demon­strat­ing the supe­ri­or­i­ty of a new mol­e­cule is dif­fi­cult. Espe­cial­ly since what is expect­ed of a new antibi­ot­ic is not nec­es­sar­i­ly to be supe­ri­or. Doc­tors would be sat­is­fied with prod­ucts of equiv­a­lent effec­tive­ness with a dif­fer­ent mech­a­nism of action.

Anoth­er dif­fi­cul­ty is that the coun­tries where the lim­its of the cur­rent phar­ma­copoeia are the most bru­tal are India and North African coun­tries. “These are coun­tries out­side Europe. Con­duct­ing clin­i­cal tri­als there is not much appre­ci­at­ed by the Amer­i­can (FDA) or Euro­pean (EMA) reg­u­la­to­ry author­i­ties,” notes Michael Mourez. And until recent­ly, the lat­ter imposed at least three pilot clin­i­cal tri­als (phase III), which requires mon­ey and many patients. “And the more peo­ple you test the prod­uct on, the more you increase the risk of tox­i­co­log­i­cal acci­dents,” he insists.

In sum­ma­ry, very few new mol­e­cules take the clas­sic devel­op­ment routem being quick­ly direct­ed towards com­pas­sion­ate use; where it is eas­i­er to obtain very ear­ly autho­ri­sa­tions for use restrict­ed to patients who have exhaust­ed oth­er ther­a­peu­tic options.

As a result, recent­ly the FDA and EMA have changed their expec­ta­tions for antibi­otics. Small­er, so-called “proof-of-con­cept” tri­als (Phase II) may now be suf­fi­cient to com­plete a mar­ket­ing autho­ri­sa­tion dossier, with restric­tions: the com­pa­ny mar­ket­ing an antibi­ot­ic in this way com­mits to fur­ther clin­i­cal tri­als as well as to increased mon­i­tor­ing of resis­tance and safe­ty of the drug (Phase IV). 

And the financial

The final prob­lem is eco­nom­ic. “The devel­op­ment of a prod­uct costs about $1.3bn and takes between 10 and 15 years,” explains Michael Mourez. “This is equiv­a­lent to a mol­e­cule in the field of immuno-inflam­ma­tion or oncol­o­gy, for exam­ple. But when a new antibi­ot­ic is approved, unlike a new anti-can­cer drug, it should not be sold, but kept for use only as a very last resort.”

“A bad antibi­ot­ic sells just as well as a good one,” he adds. So, you can’t rely on the mar­ket to reward R&D efforts. It is on the base price of the treat­ment that bio­phar­ma­ceu­ti­cal com­pa­nies expect their return on invest­ment. “Ceftolozane/tazobactam or ceftazidime/avibactam [two com­bi­na­tions launched in 2015, edi­tor’s note] cost about $10,000 per treat­ment. This sounds huge, but these are orphan dis­ease costs,” jus­ti­fies Michael Mourez. These are last-resort treat­ments, which can only be pre­scribed when first-line solu­tions have failed. “These last gen­er­a­tion com­bi­na­tions have only been sold for $100m,” adds the spe­cial­ist, “which is too lit­tle giv­en the costs of devel­op­ing and mon­i­tor­ing their mar­ket­ing, such as par­tic­i­pa­tion in the Resis­tance Obser­va­to­ry and ther­a­peu­tic edu­ca­tion programmes.” 

Com­pa­nies devel­op­ing antibi­otics are there­fore call­ing for new busi­ness mod­els. “The Beam alliance pro­pos­es patent trans­fer sys­tems to help finance small com­pa­nies that find new antibi­otics. We can also imag­ine a sort of ‘Net­flix’ of antibi­ot­ic ther­a­py, a sub­scrip­tion sys­tem for states that would ensure the financ­ing of R&D,” says Michael Mourez. Is this a start-up delir­i­um? No. The Unit­ed States has already leg­is­lat­ed on a sys­tem of this type with the Pas­teur ACT5. And the UK is con­sid­er­ing it with an eval­u­a­tion of two mol­e­cules6. Will this save our antibi­ot­ic world? Only time will tell.


Our world explained with science. Every week, in your inbox.

Get the newsletter