You are here

How to Catch COVID-19

If the last year has shown us anything, it’s that fast and accurate diagnostic tests are key for helping to control the spread of a dangerous disease. In this episode, BME’s Ashutosh Chilkoti and his PhD students explain how the lab’s signature diagnostic platform is being modified to quickly and accurately detect COVID-19.



Bells intro/music

Kane: This is Rate of Change, a podcast from Duke Engineering dedicated to the ingenious ways engineers are solving some of society’s toughest problems. I’m Michaela Kane.

This episode’s topic involves a subject that’s been hard to avoid over the last six months

****Coronavirus news coverage medley****

Back in late February and early March, most people could say that the novel coronavirus wasn’t significantly affecting their daily lives. But that wasn’t true for Cassio Fontes, a PhD student in the lab of Ashutosh Chilkoti in the biomedical engineering department at Duke University.

Chilkoti: In late February we realized that the coronavirus was here and it was going to be a serious problem, and I approached my senior graduate student Cassio Fontes and I said perhaps we should take a break, a little detour and see what we can do in terms of coronavirus diagnostics. And Cassio is very adventurous and works very hard and he jumped on it.

Fontes: I was actually home writing my thesis, right, because I’d been at Duke for quite a while and it was time to graduate, and Tosh was like ‘oh, come back, we need you to work on a test and see if you can make something, quick.’

Kane: The test Fontes is referring to is the D4 assay––the lab’s signature diagnostic platform. It’s made by inkjet-printing an array of antibodies onto a glass slide with a nonstick polymer coating, and it has the potential to transform point-of-care testing into an easy, accurate and fast process.

There are lots of different types of diagnostic tests. The gold standard is the polymerase chain reaction test, or the PCR. The PCR works by copying a target stretch of RNA or DNA until it can be detected in a sample.

This test is highly sensitive and accurate, but it isn’t portable and it requires trained technicians to use. On the opposite side of the spectrum, lateral flow tests which are similar to a pregnancy test where you can add your sample to the platform and get a diagnosis very quickly, but the trade-off is diminished sensitivity, so they aren’t as accurate.

But the D4 assay operates at the intersection of these tests, offering sensitivity and portability, and shortening a wait time of days into less than an hour.

Chilkoti and his lab began developing the platform more than 15 years earlier, but they hadn’t actually intended to create a diagnostic test at all.  Instead, it started with the creation of a polymer brush coating that prevented cells and proteins from sticking to its surface.

Kane: One of the many challenges of working with diagnostics involves cutting down on the background noise in the sample. This means that while some diagnostic tests may pick up the target protein, they may also pick up other proteins you don’t want to capture, or show a positive signal even when no target protein is present, which muddies the test result. But the non-stick coating seemed like a promising solution to this problem.

Chilkoti: And the idea, the --hypothesis was that because we have a surface to which nothing stuck accidentally when you didn’t want it to, if you attached some sort of capture molecule like an antibody to the surface, we should then be able to pull out a protein of interest from a complex mixture like blood or serum so we’d have very specific capture where we wanted it and when we wanted it to this non-stick surface.

We could really get rid of this background absorption problem that has plagued all immunodiagnostics, because a lot of diagnostics are done by putting a molecule on a surface, like an antibody, which then pulls out whatever we want and detects it. So that was our hypothesis, and we started to work on it.

I had a student Angus Hucknall, one of the most creative students I have had in my group, who discovered a way to make it work. He discovered that once you made the surface you really couldn’t change the chemistry of it. If you made any  other modifications no matter how subtle, you’d ruin this non-stick behavior.

Kane: Hucknall realized that if he dried the surface, he was able to inkjet print the antibody onto the slide so it could be captured within the polymer’s surface.

Chilkoti  -- That was a counter-intuitive idea he came up with, and by doing that he wasn’t doing any chemistry to the surface, there were no chemical modifications we made, which meant that the background and the rest of the surface remained pristine and therefore completely repellant to proteins and cells.

Kane: By figuring out how to add these capture antibodies into this polymer brush coating, Hucknall was able to create a sandwich immunoassay, which works by capturing the target protein between two antibodies, like peanut butter between two slices of bread. When a sample is placed on the slide, the antibodies capture the target antigen between them and light up. These spots of lights can then be read using a small hand-held scanner.

Chilkoti: What he showed was you can get a really intense spots, really clean spots on the background, and we showed that you could improve the performance given any sandwich immunoassay by 100 to 200 fold. That was our first development, and then in his PhD he did something equally interesting, where he decided to move into point of care diagnostics. We decided to do this because we both recognized that we were working on these surfaces where the spots that we were inkjet printing were 150 microns, which are about the width of human hair. So these were very tiny spots and we didn’t need much blood to do the assay. We could do it with a drop of blood. To turn our original test that has several steps that involve addition of reagents to the surface and wash steps, and make it as simple as a home pregnancy test meant we had to get rid of these steps.   To do this he printed little tablets of the other reagents on the chip that dissolve when blood hits the surface and then the test completes itself from that point on. It was a very simple idea but it worked beautifully. So dome years later, we published another paper in PNAS showing this idea, and that we didn’t lose any performance compared to the original test despite its extreme simplicity.

This has been a big challenge in point of care assays is that when you go from a central lab setting to a field setting or a point of care setting or a self-test setting, you often take a big hit in sensitivity, and we did not.

Kane: When Fontes joined the lab in 2014, a very different virus was drawing the world’s attention, and it prompted him to pursue the project that would later become his PhD thesis.

Fontes: 2014 was when the Ebola outbreak happened, and that’s when I jumped into this project of making a new test for a new biomarker on Ebola.

Kane: Because the Ebolavirus is only contagious after patients develop symptoms, doctors relied on sensitive tests, like PCR to identify and quarantine patients who tested positive for the virus. But Fontes and his collaborators across Duke and at a BSL4 lab in Galveston Texas, wanted to see if the D4 could detect a biomarker that was present in blood before symptoms appeared.

Fontes: The test is really easy and cheap to manufacture, so you could potentially mass-screen populations, like there was an outbreak in a village you can just screen everyone in the village and then you can actually identify people that are infected by asymptomatic or before they develop symptoms.

Kane: Although the speed, portability and sensitivity of the D4 makes it an ideal platform for tracking disease outbreaks in resource-limited settings, Fontes’s work with Ebola also proved that the D4 assay was a great platform for screening antibodies and other reagents which helped researchers identify which materials were the best match for different antigens.

Fontes: I sort of jumped into this project of making a new test for a new biomarker for Ebola, but to make that test I had to develop a bunch of different technologies to integrate new reagents, screen new reagents, develop new reagents for Ebola because it was a new biomarker and antibody development takes quite a while, and a good chunk of my work was actually figuring out how to streamline the development and selection of these antibodies, and the D4 was actually the centerpiece to expedite the antibody development.

Kane: When the coronavirus was identified as an emerging global threat in early 2020, the Chilkoti lab recognized that their previous work with the assay made it an ideal platform to test for the new public health threat.

The team was able to collect a multitude of antibodies and other reagents to test, including antibodies from the original SARS virus, which is another coronavirus that was responsible for an outbreak in 2003.

Chilkoti: We dropped everything else and started to play around with these reagents to see what kind of diagnostics we could develop.

Kane: The team was able to create a proof-of-concept that could accurately detect biomarkers from a synthetic version of the SARS-COV-2 virus, which causes COVID-19. In recognition of this work, they were awarded a RAPID Response Research grant from the National Science Foundation to continue to optimize and validate their platform. This work also involves  exploring how they can make the diagnostic easier for people to use in the field, whether that’s in places like West Africa or at the Duke University Medical Center.

Chilkoti: If you have a drop of blood diffusing on a surface, you want it to be enclosed. So let’s say you’re working with Ebolavirus. You don’t want the tester to get Ebola. It has to be completely self-contained and safe, and I have a student, David Kinnamon who has done an amazing job creating a completely self-enclosed passive microfluidic chip.

Kane: Here’s David Kinnamon to explain.

Kinnamon: Our goal has been, always, is can we add the sample and then the chip does everything else for you. So essentially what the microfluidic chip does is the incubation, the washing, and the drying all by itself. And we wanted to accomplish that in a passive way, so it doesn’t need any sort of actuators, or any sort of pumps to accomplish this task. We didn’t want any sort of secondary equipment that’s needed to operate those steps. We wanted it to be able to stand alone in the field.

Kane: As Kinnamon redesigns the microfluidic chip, they’ve also recruited Jake Heggestad, another PhD student in the Chilkoti lab, to adapt the D4 to perform both antigen and serology tests to detect COVID-19.

Chilkoti: The reagents were not great, but they were good enough for us to create a diagnostic to test for the disease, but also to develop a diagnostic to look at the antibody response. Because after you get the disease and you recover, you then develop antibodies which may be protected, and that’s called serology. And those are the two important tests. To diagnose people, or newly infected people, and also to look at and see if patients are developing antibodies, which could reflect that. We realized we could potentially do both.

Heggestad: The idea with the antigen test is that it could be used for acute diagnosis, so if someone is actively sick with the disease you could hopefully go in and get tested for these different antigens that we’re trying to detect for.

Whereas for serology test you’re trying to detect previous exposure to the disease, so you’re looking for your antibodies response to the virus, which is in the way of production of antibodies, so if you can detect antibodies to the SARS-CoV-2 virus it is suggested that you’ve had the disease at some point.

Kinnamon: For the antigen test the key thing is we don’t need a centralized lab. Like the current nasal swab involves doing the swab and it gets sent to a lab where the PCR is done, and then you get results a few days later. If our antigen test works as we think it will, we’d have quantitative results in the frame of 1-2 hours as opposed to 1-2 days depending on where those results are being sent.

Heggestad: The unique thing about our technology is that the way I like to describe it is that we have the ease of use and the usability of the lateral flow assays, which are like the pregnancy type tests that are very simple to use. But the issue with those is that they aren’t as sensitive or specific to the disease. Because of our platform we have the usability of the lateral flow assays but also the sensitivity of traditional laboratory-based assays like Elisa. A lot of that comes down to basically the surface that we’re using. It’s a high performance surface for clinical diagnostic tests.

Kane: According to the team, the goal is to test the accuracy of the platform using samples from patients, which is possible due to a collaboration with Chris Woods, a clinical investigator and Professor of Global Health and Medicine and Chief of the Infectious Diseases Division at the VA Medical Center. They anticipate starting clinical tests this fall.

Kinnamon: We have a functioning research prototype for both the antigen and the serology. I’d say the serology is further along. There are reasons why we believe it could be more useful in the short term, but our long-term goal is to have both tests working well. And when I say further along I mean we’ve tried more reagents, optimized more for performance and tested in simulated human samples for the serology tests. We’re still looking for the best reagents to use for the antigen tests. We have a functional test for the antigens but we do think we can do better.

Kane: Although the coronavirus diagnostic is a large focus of their current diagnostic work, the Chilkoti lab is also continuing to adapt their platform to explore how they can detect a variety of different diseases and pathogens. In April it was announced that the lab had received two grants totaling more than 13.7 million in funding to use the platform to detect bioterrorism agents and to diagnose breast cancer in resource limited settings.

After all, the current global pandemic underscores the need for rapid, affordable and accurate diagnostic testing.

Fontes: The platform itself is not only sensitive, but it’s also adaptable to track a lot of diseases. Like we’ve done cancer biomarkers, we’ve done several infectious diseases, we’ve done inflammatory biomarkers. We can potentially in the future, as the technology develops, we can make a panel for I don’t know, 50 diseases in this platform. All you need is a little drop of blood and you can test for all of them at once.

Testing is the first step to figure out who is sick, who is not sick, to isolate people, to start decontamination. Even to treat people. And if you’re not testing people and if you can’t test people fast enough, there is really no way to control any outbreak.

Kane: That’s it for this week’s episode. If you liked our podcast, share with your friends, follow us on social media for updates, and don’t forget to subscribe! Thanks for listening to Rate of Change, and until next time, stay safe!