When scientists concluded that the AstraZeneca and Johnson & Johnson vaccines were associated with extremely rare blood clots, they had only the beginnings of a mystery. What in the vaccine was causing the clots, and why were the clots only happening to a few select people?
Now, scientists are starting to piece together a series of clues to find the answer. It started with a Tucson teenager who died more than 20 years before this vaccine was created. Then there was a blood-thinning drug that had occasionally caused similar types of clotting issues. And now there are images from a computer modeling system in Tempe so precise it has to be insulated from the vibrations of the earth.
A new article released in the journal Science Advances, which includes significant work from Arizona researchers, has unlocked part of the puzzle of why certain people had complications from adenovirus-based COVID-19 vaccines such as Johnson & Johnson’s. With data from across disciplines compiled and tested at a pace never seen before the pandemic, scientists say they have identified some of the key mechanisms behind the proteins and molecules in our vaccines and our bodies.
This research is only the beginning: Building on the answers they found here could have implications for the development of future vaccines as well as therapeutics for cancers and other genetic disorders. And one day, people may be able to take a test pre-vaccine to determine their risk of developing a blood clot.
“The science doesn’t stop here with trying to fight the pandemic,” said Abhishek Singharoy, one of the co-authors of the paper. “Now, based on this data, they may be able to improve the vaccine for (the) next generation.”
Before scientists could even begin trying to figure out the mechanisms behind the formation of these blood clots, they first had to be able to see what they were working with.
Singharoy has been working on that for years.
Singharoy, an assistant professor in ASU’s School of Molecular Sciences and the Center for Applied Structural Discovery at the Biodesign Institute, works with a “computational microscope.” Essentially, he helps develop the computer code for a virtual 3D Lego set that scientists can use to build models of what goes on inside our body at a molecular level.
In the basement of a lab in Tempe, on top of a four-foot slab of reinforced concrete that keeps any minuscule seismic vibrations from shaking the equipment, researchers use a high-tech microscope — the only one of its kind in Arizona — to take thousands of two-dimensional images of molecules from different angles. Then biophysicists like Singharoy can reconstruct a virtual 3D model in a process similar to the one used to create some 3D movies.
Once they have the computer models up and running, scientists can simulate interactions between different cells and proteins in the body to get a better sense of which experiments might be promising in the lab.
Clinical data had already shown that the blood clots associated with the AstraZeneca and Johnson & Johnson vaccines resembled rare blood clots that occurred in some patients being treated with an anti-clotting medication called heparin, according to a news feature published by the journal Nature in August. In those cases, negatively charged heparin molecules bind to a positively charged protein called platelet factor four (PF4), which is found in all healthy humans.
Heparin and PF4 attract each other like magnets. Was something in the AstraZeneca and Johnson & Johnson vaccines acting like a magnet for PF4, too?
Enter Singharoy’s simulations. The team modeled the proteins on the outer shell of the virus that encases the genetic materials of the vaccine, as well as the PF4 protein. Then they put them together in a computer simulation.
Sure enough, the adenovirus shell and PF4 were attracted to each other.
“I think this is the first time we could see a direct application of molecular simulation into an industrial-scale application,” Singharoy said. “Normally these things happen in serial: I do a model, then people do experiments. But this time around, it’s literally overlapping. Whatever we’re doing essentially gets tested in a matter of weeks and not months or years. So that’s a huge amount of responsibility on a model.”
Once the researchers could see the substances of interest attracted to each other on the computer, they had to see if it was happening in real life, too.
They do that using “a very sophisticated biophysical technique,” explained Alexander Baker, a research fellow at the Mayo Clinic Center for Individualized Medicine in Scottsdale. Baker is a viral vector engineer, which means he finds ways of “making viruses do what we want them to do,” from developing new vaccines to targeting cancers.
Baker and his team collaborated with an AstraZeneca team to use what the experts call a microfluidic device, which electrochemically coats tiny glass tubes with any molecules researchers want. In this case, they used a coating of the viral shell used in the adenovirus vaccines.
Next, scientists can flush the tube with whatever they want to test: In this case, a saline solution containing PF4, to see if anything sticks. If it does, they can prove PF4 is attracted to the viral shell, just as it’s attracted to heparin.
And that’s exactly what they found. Compared to a control without the adenovirus on the tubes, where they saw no “stickiness,” the viral shell and PF4 showed more affinity for each other, Baker said.
Singharoy and Baker both described how their research is the result of specialists working together, using a broad range of techniques. Baker even admitted that he doesn’t fully understand the mechanics of how microfluidic devices work because they require so much highly specific knowledge from a variety of different fields.
“These devices are made by engineers and physicists in collaboration with biologists,” Baker said.
That collaboration led to more than one discovery. Baker and a team at Cardiff University also conducted similar tests with the spike protein of the adenovirus and a receptor called CAR, which is how the adenovirus enters human cells. As expected, they found that they stuck to one another.
But since CAR is also found on platelets, they realized they might have yet another clue, this time to pass on to scientists who specifically study the blood.
The mechanisms of blood clots
Over 20 years ago, another adenovirus — one with the potential for an important medical breakthrough — was involved in a clinical trial gone wrong. But that trial later sparked the research that led to surprising connections between adenoviruses and blood.
In 1999, a teenager from Tucson named Jesse Gelsinger became one of the first participants in a new gene therapy trial that featured an adenovirus similar to the one used in the AstraZeneca and Johnson & Johnson vaccines. Gelsinger had a rare metabolic disorder, and the hope was that researchers could deliver genetic instructions in the vehicle of the adenovirus to help repair some of his dysfunctional genes.
Instead, Gelsinger had a severe inflammatory response and eventually died, and many gene therapy projects using adenoviruses were abandoned.
But by 2004, Dr. Maha Othman was becoming fascinated by the promise of adenoviruses as a tool for gene therapy. “(Adenoviruses) can go into cells really easily. You can manipulate the genome and make it disabled so it doesn’t infect people. (And) the genome is well studied, extensively studied. So it’s the perfect delivery vehicle,” Othman said.
Othman, now a professor at Queen’s University in Ontario, Canada, is a hematologist who focuses on bleeding and clotting disorders. Fifteen years ago, she started wondering whether there was something happening between adenoviruses and platelets that were leading to low platelet counts after gene therapy treatments, a result Othman described in a 2006 article as “consistently reported.”
When platelets are doing what they’re supposed to do, they’re a key part of your body’s toolkit to repair damage. When platelets are “activated” — say, if you get a cut and start bleeding — they trigger coagulation and eventually initiate the creation of a kind of molecular “mesh” to seal off the site of injury, so you don’t bleed out.
Normally, though, platelets aren’t activated: they’re just floating around in your bloodstream. It takes a specific signal for the platelets to realize they need to change shape and initiate the process known as hemostasis or slowing blood flow and beginning clotting.
Othman put an adenovirus (a different strain from the one used in some COVID-19 vaccines, but a biologically similar one) in a test tube with blood so they could study the platelets.
She found that the platelets were activated almost immediately. When she tested it in live mice, it took a few hours, but the same thing happened.
Othman said that so many years later, she didn’t connect the dots to the new vaccines until the rare blood clotting was reported in Johnson & Johnson as well as AstraZeneca recipients. “I thought, ‘both of them (are) adenovirus (vaccines) … This must be the platelet,” Othman said.
Although she and other hematologists still don’t fully understand why only a few select individuals experience this phenomenon, they can make some educated guesses. Together with Baker and some of their other colleagues, Othman proposed seven possible mechanisms behind the rare blood clotting. And because scientists now know that the adenovirus in the vaccines sticks to CAR and PF4, Othman favors two of those hypotheses, which she believes are part of the same story.
According to one of her hypotheses, when you get a shot, a bit of the adenovirus is able to get into your bloodstream. That virus can then stick to the CAR receptors on your platelets, causing them to become activated and release PF4. The activation in itself might be causing blood clots where they aren’t supposed to be, or there could be a few more steps involved.
Once PF4 is out and about, the adenovirus can stick to that, too. As Singharoy describes it, the vaccine is like a car that contains genetic information inside, but if PF4 gets attracted to it and hops on the outside, it can end up in the lymph nodes. That’s where some of your immune cells are hanging out, and some people’s immune cells don’t like seeing PF4 taking a joyride.
In those individuals, the immune cells start producing anti-PF4 antibodies, and eventually, the whole mess clumps together in a kind of molecular chain. If you get too many in one place, that, too, might trigger your platelets to create a clot.
Othman said it’s still very much a mystery why so few people have this response. That could be because people have such vastly different immune systems and responses, as well as different platelets. Some platelets have a lot of sticky CAR on their outsides, while others don’t have much at all. And some people have those anti-PF4 antibodies, while other people don’t. But even if they do, most people who have those antibodies still won’t develop blood clots.
“That’s the million-dollar question: Why is this not happening in everyone?” Othman asked. She said she is fascinated by the ways that the questions she was asking fifteen years ago are still being answered today — and how critical those questions have become in a pandemic.
“When science goes after an important question, it can last, and last for (a long time),” she said.
Tomorrow’s vaccine safety?
While the scientists asking these questions have more work to do, the adenovirus vaccines still remain some of the safest in the world. Only about four in 1 million adults have experienced the types of blood clots associated with these adenovirus vaccines; your chance of getting struck by lightning is about two in 1 million.
In addition, COVID-19 itself causes blood clots, and unvaccinated individuals have a much higher likelihood of experiencing them if they get sick.
Still, Singharoy expressed appreciation for his colleagues, who he says are devoted to making even more precise and safer vaccines. “Scientists really want to learn from the rarest of mistakes and want to move (the technology) forward,” he said.
AstraZeneca, which contributed to the paper by Baker, Singharoy and their colleagues, said in a statement that “it is important to note that the mechanism identified does not demonstrate that it is the cause of TTS (the disorder that causes blood clotting) and that most individuals that will have PF4 antibodies will not develop TTS.”
Singharoy agrees that there is more work to be done to understand the processes they identified, but believes that if their hypotheses are proven, further research could pave the way for new and individualized tests for vaccine safety.
“If this mechanism was established, then one can test for these antibodies,” he said, noting that while we already have the technology to test for these antibodies as a clinical tool, antibody testing for everyone who wants to get vaccinated could be expensive. Baker agreed, especially in comparison to how extremely rare these blood clots are.
But someday, after many more years of research, Singharoy thinks maybe a test could be an extra precautionary measure. If that happened, “for people that find these antibodies…we would not use (AstraZeneca or Johnson & Johnson vaccines),” he said.
In the meantime, Othman sees the research as a win for collaboration among scientists with different specialties. “(For) many of the publications that come out (on this) topic, you will see 10, 15 different people on it,” she said. “Everybody has an expertise in some aspect and contributes to that understanding.”
Independent coverage of bioscience in Arizona is supported by a grant from the Flinn Foundation.
Melina Walling is a bioscience reporter who covers COVID-19, health, technology, agriculture and the environment. You can contact her via email at firstname.lastname@example.org or on Twitter @MelinaWalling.