Covid, pregnancy, heart attack: how to improve tests

Since the Cov­id-19 pan­dem­ic, we have all become famil­i­ar with lat­er­al flow tests (LFA), com­monly referred to as “anti­gen tests”. This type of test can, in prin­ciple, detect the pres­ence of any pro­tein anti­gen in body flu­ids (urine, blood, saliva etc.) – not only SARS-CoV­‑2. Their advant­age over con­ven­tion­al tests per­formed in med­ic­al labor­at­or­ies (such as ELISA) is that they are fast (15 minutes instead of sev­er­al hours), inex­pens­ive and do not require heavy labor­at­ory equipment. 

They are there­fore use­ful for detect­ing many dis­eases, but also changes in the body such as preg­nancy, or the pres­ence of a pro­tein food that could cause food pois­on­ing. How­ever, unlike ELISAs, one of the lim­it­a­tions of LFA tests is that they only give a qual­it­at­ive answer. For a preg­nancy test, for example, the res­ults are still lim­ited to “yes/no” – either you are preg­nant or you are not. How­ever, there are many cases where it would be use­ful to determ­ine the amount of the pro­tein being tested for. For example: “yes” you are preg­nant, but for how many weeks? Or, “yes” you have Cov­id, but with what vir­al load? 

Our meth­od can be con­sidered 40 to 50 times more sens­it­ive than tra­di­tion­al ones.

Anoth­er lim­it­a­tion of LFA tests is that they are not par­tic­u­larly sens­it­ive. Devel­op­ing quant­it­at­ive and sens­it­ive LFA tests would push the bound­ar­ies of what these tests can be used for. They could even indic­ate how long ago a preg­nancy star­ted. This sens­it­iv­ity and accur­acy would also allow the test to detect the amount of cTnI pro­tein nat­ur­ally secreted before car­di­ac arrest, for example. Coupled with its new access­ib­il­ity, it could be per­formed dir­ectly in the ambu­lance, facil­it­at­ing patient management.

Fanny Mousseau, a research­er at the Optics and Bios­ciences Labor­at­ory at Insti­tut Poly­tech­nique de Par­is, is work­ing on the devel­op­ment of LFA tests based on a new tech­no­logy. LFA tests are based on the use of detec­tion anti­bod­ies that attach to the tar­geted pro­teins. “The anti­body (~10 nm) in this type of test must be coupled to some­thing per­cept­ible to the naked eye,” explains the research­er, “some­thing lar­ger and col­oured, hence the use of gold nan­o­particles (~50–100 nm) in com­mer­cial tests. The first step in our work was there­fore to find a sub­sti­tute with a more intense and quan­ti­fi­able optic­al signal.”

The innov­a­tion: repla­cing gold nan­o­particles with lumin­es­cent nan­o­particles. “By using lumin­es­cent nan­o­particles, we were able to improve the sens­it­iv­ity of the test and make it quant­it­at­ive and more robust.” And after some research into optim­ising the use of these nan­o­particles, the Laboratory’s team man­aged to use 4x few­er of them than gold nan­o­particles, which amounts to using 120x less detec­tion anti­body and thus con­sid­er­ably redu­cing the cost of a test. “When ana­lys­ing LFAs made with our nan­o­particles with the naked eye, the low­est detect­able pro­tein con­cen­tra­tion is 5x to 10x lower,” says the research­er. “Using our optic­al ana­lys­is meth­od, made with the help of the applic­a­tion we developed for this pur­pose, we can detect con­cen­tra­tions that are 2–4 times lower. With the two com­bined, our meth­od can be con­sidered 40–50 times more sens­it­ive than the tra­di­tion­al one.”

In addi­tion to this increase in sens­it­iv­ity, using lumin­es­cent nan­o­particles allows this test to become quant­it­at­ive, instead of qual­it­at­ive as with gold nan­o­particles. “A test is con­sidered accur­ate if after three repe­ti­tions we get the same meas­ure­ment three times,” says Fanny Mousseau. “How­ever, by accur­ately determ­in­ing the quant­ity of tar­geted pro­teins present in the body, we can, for example, determ­ine the stage of a dis­ease and fol­low its evolution.”

The Laboratory’s team is also work­ing on the poten­tial that this meth­od has with regard to mul­ti­plex­ing – the pos­sib­il­ity of detect­ing sev­er­al pro­teins in a single test, and there­fore in a single pro­ced­ure. “In the case of cer­tain patho­lo­gies, it is not the evol­u­tion of a single pro­tein that is inter­est­ing, but that of sev­er­al,” explains the research­er. For example, endocan is a bio­mark­er of inflam­ma­tion. This pro­tein exists in two forms, ‘nat­ive’ and ‘cleaved’, and know­ing the respect­ive quant­it­ies of these two forms may make it pos­sible to bet­ter define the treat­ment of cer­tain pul­mon­ary dis­eases. To decide on the most suit­able treat­ment for the patient accord­ing to his oth­er con­di­tions, it is neces­sary to observe the quant­ity of “nat­ive” endocan and that of “trans­formed” endocan.

We are able to determ­ine the pres­ence of three dif­fer­ent pro­teins in one mul­ti­plexed test with an accur­acy rate of 30%.

Devel­op­ing this poten­tial is still a work in pro­gress: at the moment the research team does not con­sider this meth­od to be reli­able enough for mul­ti­plex­ing. The reas­ons for this lack of reli­ab­il­ity are known, how­ever, and are due to the phe­nomen­on of cross-react­iv­ity. “For the detec­tion of two pro­teins, two sep­ar­ate detec­tion sys­tems are needed [one for the anti­body, anoth­er for the nan­o­particles], which gives rise to the pos­sib­il­ity of so-called cross-react­iv­ity between the dif­fer­ent systems.”

“So far, mul­ti­plexed tests are still qual­it­at­ive. This year, I developed an innov­at­ive cal­ib­ra­tion curve, which takes this cross-react­iv­ity into account, to try to read the res­ults of a mul­ti­plex more accur­ately. As a res­ult, we can determ­ine the pres­ence of three dif­fer­ent pro­teins in one mul­ti­plexed test with an accur­acy rate of 30%. This is a prom­ising start,” she says. The research­er is optim­ist­ic that clin­ic­al tri­als will begin soon. 

Once the tech­nique for improv­ing the sens­it­iv­ity of LFA test strips had been iden­ti­fied, all the research­ers had to do was devel­op an ana­lys­is meth­od for these tests. “The main applic­a­tion of our research was to replace blood sampling,” says Fanny Mousseau. With our meth­od, the res­ults are quick­er, but just as accur­ate. This meth­od was developed with a view to facil­it­at­ing these ana­lyses in places where access to blood tests is lim­ited [in war zones, in devel­op­ing coun­tries, etc.], or when they are urgently needed [in an ambu­lance in case of a heart attack, for example].

As a res­ult, the ongo­ing research required a tool to illu­min­ate the nan­o­particles to make it easi­er to read the res­ults. “The ana­lys­is meth­ods required the use of huge equip­ment, which was con­tra­dict­ory to the port­able advant­age of this type of test,” explains the research­er. “In our labor­at­ory, we inven­ted a small device cap­able of per­form­ing the test. This small device, which is thick­er but not lar­ger than a mobile phone (~10 cm x 5 cm x 5 cm), comes with an applic­a­tion with which the user pho­to­graphs the test.”

“We have integ­rated the ana­lys­is pro­gram. The user simply takes the pic­ture and presses the con­trol strip on the LFA strip (the strip that veri­fies that the test has worked). The algorithm, know­ing the exact dis­tance between the con­trol strip and the test strip (the strip on which our pro­tein of interest is detec­ted) will ana­lyse, pixel by pixel, the col­our intens­ity of the test strip.

The res­ult is then dis­played, simply indic­at­ing the amount of tar­get pro­tein present in the test. “In the long term, the goal is to sim­pli­fy the pro­cess as much as pos­sible so that every­one can have access to it,” she concludes. 

Pablo Andres